Irving Wladawsky-Berger

The Potential of AI Agents: Three Stories Drawn from Real-World Experiences

About three years ago I started following and writing about Generative AI (GenAI), Large Language Models (LLMs), chatbots, and related topics. Then a few months ago I started hearing about agentic AI, and since then I’ve been trying to better understand what agentic AI is all about.

I asked Babson College professor Tom Davenport, a prolific writer whom I have long known, if he has written any recent articles on agentic AI. Davenport recommended Agentic Artificial Intelligence: Harnessing AI Agents to Reinvent Business, Work and Life, — a book published in March of 2025 that was the result of a unique collaboration among a diverse team of 27 professionals in business, academia, programming, and research, — where Davenport was one of the principal authors.

As we can expect given the number of co-authors, this is a long, comprehensive book at over 550 pages and 14 chapters. I found the book’s Introduction chapter to be a very good overview of the value of Agentic AI, particularly its explanation of the difference between Agentic and Generative AI. Let me summarize what I learned from the Introduction.

Agent Explainability: The Foundation for Trust in the Agent Ecosystem

What is Agentic AI? There are multiple answers to this question, not surprisingly since we’re still in the early stages of this complex technology. Wikipedia emphasizes the autonomous nature of agentic AI in its definition: Agentic AI is “a class of artificial intelligence that focuses on autonomous systems that can make decisions and perform tasks without human intervention.”

McKinsey noted that AI is now “moving from thought to action” in a July, 2024 report that explained the difference between generative and agentic AI. “We are beginning an evolution from knowledge-based, gen-AI-powered tools — say, chatbots that answer questions and generate content — to gen AI-enabled agents that use foundation models to execute complex, multistep workflows across a digital world. … Gen AI agents eventually could act as skilled virtual coworkers, working with humans in a seamless and natural manner.”

Similarly, IBM answers the question “What is Agentic AI” by writing: “Agentic systems provide the flexibility of LLMs, which can generate responses or actions based on nuanced, context-dependent understanding, with the structured, deterministic and reliable features of traditional programming. This approach allows agents to think and do in a more human-like fashion.”

The Future of Agentic AI Ecosystems

Wikipedia defines Agentic AI as “a class of artificial intelligence that focuses on autonomous systems that can make decisions and perform tasks without human intervention.” A recent IBM website, “What is Agentic AI,” adds that: “Agentic systems provide the flexibility of LLMs, which can generate responses or actions based on nuanced, context-dependent understanding, with the structured, deterministic and reliable features of traditional programming.”

“Broadly speaking, agentic systems refer to digital systems that can independently interact in a dynamic world,” explained a McKinsey report published in July of 2024.  “While versions of these software systems have existed for years, the natural-language capabilities of gen AI unveil new possibilities, enabling systems that can plan their actions, use online tools to complete those tasks, collaborate with other agents and people, and learn to improve their performance. Gen AI agents eventually could act as skilled virtual coworkers, working with humans in a seamless and natural manner. … In short, the technology is moving from thought to action.”

“Billions of dollars of investment by some of the largest firms on the planet are flowing into tools that will make it easy to build autonomous agents,” wrote senior technology executive Eric Broda in “Agentic Mesh: The Future of Generative AI-Enabled Autonomous Agent Ecosystems,” a November 2024 article in the Medium platform. “And if this huge investment, and the recent headlines, are any indication, we will soon have many, many autonomous agents collaborating in a dynamic ecosystem.”

Is AI a Normal Technology?

Hard Fork is a weekly podcast from the NY Times that examines the impact of the latest developments in technology on the economy and society. The podcast is hosted by Kevin Roose, NY Times technology journalist, and Casey Newton, founder of the Platformer newsletter. A few weeks ago the podcast discussed Is AI a ‘Normal' Technology? with Princeton computer science professor Arvind Narayanan, co-author of AI Snake Oil: What Artificial Intelligence Can Do, What It Can’t, and How to Tell the Difference with Princeton PhD candidate Sashay Kapoor, a book that was published in September of 2024.

Roose and Newton started the podcast by noting that the previous week their guest at the podcast had been AI researcher Daniel Kokotailo, founder and executive director of the AI Futures Project, where they discussed their new article, AI 2027, which predicted that “the impact of superhuman AI over the next decade will be enormous, exceeding that of the Industrial Revolution.” Narayanan is highly skeptical of the fast takeoff predictions of AI 2027. In a recent essay also co-authored with Kapoor, “AI as a Normal Technology,” they laid out an alternative vision of AI, “basically one that treats AI not as some looming superintelligence that’s going to go rogue and take over humanity, but as a type of technology like any other, — like electricity, like the internet, like the PC, — that have taken a period of years or even decades to fully diffuse throughout society.”

This is a very important question I’ve been thinking a lot about. Will AI’s impact on economies and societies be similar to those of previous technological revolutions of the past two and a half centuries, — that is, Is AI a normal technology?, — or is AI in a class by itself, destined to exceed the impact of the Industrial Revolution in just one decade?

The 2025 AI Index Report: AI’s Influence Across Society Continues to Transform the Way We Live and Work

The 2025 AI Index Report was released in early April by the Stanford Institute for Human-Centered Artificial Intelligence, — it’s eigth annual analysis of the impact, progress, and trends of AI. Led by an interdisciplinary group of experts from across academia and industry, the AI Index report aims to be the world’s most authoritative source for data and insights about AI.

“What a year 2024 has been for AI,” wrote the report’s co-directors Yolanda Gil and Raymond Perrault. “The recognition of AI’s role in advancing humanity’s knowledge is reflected in Nobel prizes in physics and chemistry, and the Turing award for foundational work in reinforcement learning. The once-formidable Turing Test is no longer considered an ambitious goal, having been surpassed by today’s sophisticated systems. Meanwhile, AI adoption has accelerated at an unprecedented rate, as millions of people are now using AI on a regular basis both for their professional work and leisure activities. As high-performing, low-cost, and openly available models proliferate, AI’s accessibility and impact are set to expand even further.”

“AI is no longer just a story of what’s possible — it’s a story of what’s happening now and how we are collectively shaping the future of humanity,” Gil and Perrault added. At over 450 pages and eight chapters, the 2025 AI Index is recognized globally as one of the most comprehensive and authoritative resources on artificial intelligence, referenced in hundreds of academic papers and used by policymakers and government agencies around the world.

Pathways and Challenges Toward an Open Data Ecosystem

“Our data is everywhere and powering everything,” noted “Pathways to Open Data,” a report by Linux Foundation Research published in March of 2025. “From marketing, to healthcare, to government services, to the emerging phenomenon of programming AI agents, organizations leverage data to be as efficient and effective as possible. However, data is often siloed within entities and any third-party data access requires overcoming significant technical, legal, economic, operational, and cultural obstacles that are multifactorial and at times may seem intractable. The increasing reliance on data calls for an assessment of these obstacles and how organizations can shift toward greater openness and sharing.”

The report defines open data as “data infrastructure that has the technical and legal requirements in place to make the data freely accessible for universal use, reuse, and redistribution.” Open data has its roots in open science and open knowledge in general, where non-personal and non-commercial data is freely published for the purpose of greater innovation, transparency, and collaboration. This culture of openness is strongest in public institutions, where the data collected is considered a public good without profit-generating opportunities, and where transparency of government and public-sector information is encouraged.”

Many governments have developed open data portals to enable access to public information, such as the Public Sector Information Directive launched by the European Union in 2003, and Data.Gov, launched by the Obama administration in May of 2009 with the goal of increasing transparency and accountability through open data. Initially launched with 47 data sets, Data.Gov now includes over 300,000 open data sets.

What Are AI Agents and How Should We Use Them?

After decades of promise and hype, AI has (finally) come of age. The 2010s saw increasingly powerful deep learning systems surpass human levels of performance in tasks like image and speech recognition, skin and breast cancer detection, and winning at championship-level Go. The 2020s saw the advent of generative AI systems, with their impressive ability to create high-quality original content, — e.g., text, images, video, audio or software code, — in response to a a user’s natural language prompts. And now, we’re seeing the emergence of agentic AI.

Wikipedia defines agentic AI as autonomous AI systems “that can make decisions and perform tasks without human intervention.” Agentic AI systems are designed to “autonomously make decisions and act, with the ability to pursue complex goals with limited supervision. It brings together the flexible characteristics of large language models (LLMs) with the accuracy of traditional programming.”

A few weeks ago I posted a blog, Agentic AI: The Evolution of Application Development, mostly based on a McKinsey Digital report, “Why agents are the next frontier of generative AI.” Agentic AI systems are now taking AI to the next level, enabling us to automate processes that can plan and execute their actions with limited human intervention. “In short, the technology is moving from thought to action.”

Informational GPS: Navigating the Complex World of 21st-Century Knowledge

“Until the early 2000s, paper maps did not exist — they were simply called maps” wrote Reid Hoffman and Greg Beato in “Informational GPS,” one of 12 essays in The Digitalist Papers, a roadmap of the possible futures that the AI revolution might produce. Before May of 2000, the Global Positioning System (GPS) was deliberately limited for civilian use due to security concerns. But then access to GPS was made available to everyone as a public good

“In the wake of this policy shift, GPS experienced a Cambrian explosion of innovation.” Improved performance and falling prices for consumer receivers quickly turned GPS into an indispensable part of life in the 21st century. A 2019 study on the Economic Benefits of GPS by the National Institute of Standards and Technology (NIST) estimated that by 2017, GPS had generated roughly $1.4 trillion in location-based services across a variety of industries since being made available for civilian and commercial use.

The term GPS has entered the popular vernacular to mean one’s specific location at a specific point in time. Its breakthrough application has been turn-by-turn navigation, enabling us to move through the physical world with precise, constantly updated information. “At literally every turn, these navigation systems increase individual agency by telling us where we are, what else is nearby, what obstacles might impede our progress, and so much more.”

A major reason why GPS has been so widely adopted is that it offers us individual agency along with broad access to navigation information. GPS not only suggests the best routes for us to follow based on current traffic conditions, but it also offers us the ability to modify our route along the way and quickly readjusts to accommodate our wishes and new position.

Agentic AI: The Evolution of Application Development

I’ve been following the evolution of AI since the 1970s, especially the more recent era of data-centric AI systems based on highly sophisticated models trained with large amounts of information and powerful computer technologies. We were wowed when in 1997 Deep Blue won a celebrated chess match against then reigning champion Gary Kasparov, — one of the earliest, most concrete grand challenges of AI. The 2010s saw increasingly powerful deep learning AI systems surpass human levels of performance in a number of tasks  like image and speech recognition, skin and breast cancer detection, and playing championship-level Go.

More recently, the impressive ability of large language models (LLMs) and chatbots to interact with us and generate cogent, articulate sentences has given us the illusion that we’re dealing with a well educated, intelligent human, rather than with a sophisticated stochastic parrot that’s been trained with huge amounts of human language but has no human-like understanding of the ideas underlying the sentences it’s putting together.

Agentic systems are now taking AI to the next level.

A Planetwide Vision of the Future of Global Supply Chains

“The central purpose of supply chains is to fulfill the needs and wants of humanity — delivering all the food, medicine, energy, apparel, and other worldly goods needed by the eight billion people in the planet,” wrote MIT professor Yossi Sheffi in the preface of his 2023 book, The Magic Conveyor Belt: Supply Chains, A.I. and the Future of Work. “The societal and economic spasms of the early 2020s highlighted the crucial role of world-spanning supply chains in the modern global economy, as well as the growing role of digital technology, including A.I. and automation, in the future economy.”

I’ve been following the evolution of global supply chains for the past three decades. The 1990s ushered a golden age of globalization, when the world seemed to be coming together. Thomas Friedman’s The World is Flat became an international best-seller in 2005, nicely explaining what globalization was all about, including the key forces that contributed to the flattening of the world, from the explosive growth of the Internet to the rise of just-in-time global supply chains.

My interest in global supply chains increased significantly in the mid 2010s, as a result of my growing interest in the business applications of blockchain technologies. In 2016, the World Economic Forum (WEF) included blockchain in its annual list of Top Emerging Technologies. The WEF report compared blockchain to the Internet, noting that “Like the Internet, the blockchain is an open, global infrastructure upon which other technologies and applications can be built. And like the Internet, it allows people to bypass traditional intermediaries in their dealings with each other, thereby lowering or even eliminating transaction costs.”

I became convinced that global supply chains would emerge as the killer app for enterprise blockchain applications. Over the following years, I closely tracked the evolution of blockchain-based supply chain systems. I assumed that, much like the Internet in the 1990s and Linux in the 2000s, it was just a matter of time before blockchain technologies would be widely embraced in the business world.

But, as we now know, things didn’t quite turn out as I had expected. “Blockchain, the technology underpinning bitcoin and other cryptocurrencies, for years has been viewed by some companies as a way to drive industry-transforming projects, among them the tracking of assets through complex supply chains. So far that hasn’t happened,” wrote the Wall Street Journal in a December 2022 article, “Blockchain Fails to Gain Traction in the Enterprise.”

Business Strategy in the Emerging AI Era

“AI is changing the cost and availability of expertise, and that will fundamentally alter how businesses organize and compete,” said “Strategy in an Era of Abundant Expertise,” a Harvard Business Rreview (HBR) article by Harvard Business School professor Karim R. Lakhani and Microsoft corporate strategists Bobby Yerramilli-Rao, John Corwin, and Yang Li. At its most basic level, a business can be thought of as a bundle of expertise organized to accomplish specific tasks for competitive advantage. The evolution of this expertise defines the evolution of the business.

The article defines expertise as “a combination of deep theoretical knowledge and practical know-how in a specific domain.” Expertise can take many forms, such as deep subject matter knowledge in a specific field, — like manufacturing, software engineering, or a medical specialty, — and the skills needed to successfully run a business, — e.g., marketing, sales, operations, finance. “Companies create  value by applying their expertise efficiently at scale to solve problems for their customers. Typically they possess it in a variety of areas, but most differentiate themselves through their unique proficiency in just a handful of activities that are fundamental to how they create competitive advantage.”

While critical to a company’s success, keeping up with technological advances can be quite challenging for two key reasons. First, expertise is constantly expanding, making it difficult to stay abreast of advances and insights in a competitive global marketplace. And second, the cost of accessing that expertise is constantly falling, lowering the barrier for new, fast moving competitors. The interplay between these two forces, — “the increasing amount of expertise required to create value and the decreasing cost of accessing that expertise — shapes companies and affects the scope of their operations.”

While AI Has Been Eating Computing, Philosophy May Be Ready to Make a Meal of AI

In his legendary 1965 paper, “The Future of Integrated Circuits,” Intel co-founder Gordon Moore first made the empirical observation that the number of components in integrated circuits had been doubling every year since their invention in 1958. Moore predicted that the trend would continue for at least ten years, a prediction he subsequently changed to a doubling every two years. The semi-log graphs associated with Moore’s Law have since become a visual metaphor for the revolution unleashed by the exponential improvements of just about all digital components, from processing speeds and storage capacity to networking bandwidth and pixels.

Moore’s Law has had quite a run, but like all things based on exponential improvements, it eventually slowed down and flattened out in the 2010s. What will now drive computing in the post-Moore’s-Law era?

Software has been a key driver of computing since the 1980s. In a 2011 essay, “Software is Eating the World,” technologist and investor Marc Andreessen wrote that “we are in the middle of a dramatic and broad technological and economic shift in which software companies are poised to take over large swathes of the economy. More and more major businesses and industries are being run on software and delivered as online services — from movies to agriculture to national defense.” Entrepreneurial companies all over the world have been disrupting established industries, with innovative software-based solutions.

In 2017, Nvidia CEO and co-founder Jensen Huang boldly updated Andreessen’s essay by predicting that “Software Is Eating the World but AI is Going to Eat Software.” In 2024, he further predicted that “the future of coding as a career might already be dead in the water with the imminent prevalence of AI.

“But even as software eats the world and AI gobbles up software, what disrupter appears ready to make a meal of AI?,” asked MIT researchers Michael Schrage and David Kiron in “Philosophy Eats AI,” a recent article in MITSloan Management Review. “The answer is hiding in plain sight. It challenges business and technology leaders alike to rethink their investment in and relationship with artificial intelligence. There is no escaping this disrupter; it infiltrates the training sets and neural nets of every large language model (LLM) worldwide.”

The Impact of AI on Research and Innovation

Postscript: On May 16, 2025, MIT issued a press release explaining the concerns raised about the integrity of the research in the preprint of the paper “Artificial Intelligence, Scientific Discovery, and Product Innovation” that was published on arXiv in November 2024 and referenced in this blog.

=======================================================================================

On December 29, the WSJ  published “Will AI Help or Hurt Workers?,” an article based on a research
paper by Aidan Toner-Rodgers, a second year PhD student in MIT’s Economics Department.

One of the reasons the WSJ article caught my attention is that it featured a photo of the MIT graduate student in between two of the world’s top economists whose research I’ve closely followed for years: Daron Acemoglu, — who in October was named a co-receipient of the 2024 Nobel Memorial Prize in Economic Science, and David Autor (along with his dog Shelby) — who was a co-chair of a multi-year, MIT-wide Taskforce on the impact of AI on “The Work of the Future.”

Both professors raved about Toner-Rodger’s research even though they have somewhat different points of view on the impact of AI on workers. Professor Autor is more optimistic, arguing that “AI Could Actually Help Rebuild The Middle Class.” His friend and colleague, professor Acemoglu, worries that AI could actually worsen income inequality and not do all that much for productivity and GDP growth over the coming decade. “For all the talk about artificial intelligence upending the world, its economic effects remain uncertain,” he said in a recent interview. “There is massive investment in AI but little clarity about what it will produce.” But they both agreed that “the research by Toner-Rodgers, 26 years old, is a step toward figuring out what AI might do to the workforce, by examining AI’s effect in the real world.”

Leveraging GenAI for Workforce Productivity: a Scenario Based Analysis

“Generative artificial intelligence (GenAI) has the potential to drive significant improvements in workforce productivity at the level of tasks, organizations and economies,” said a recent report “Leveraging Generative AI for Job Augmentation and Workforce Productivity: Scenarios, Case Studies and a Framework for Action,” by the World Economic Forum in collaboration with PwC. “Delivering those gains depends, among other things, on the deployment of GenAI to augment jobs, i.e. to partially perform tasks in such a way that technology effectively supports or enhances human capabilities through human-machine collaboration.”

“With such a fast-moving technology, it is hard to predict how even the relatively near-term future will play out,” the report added. “To help think through the possibilities, it is useful to think in terms of scenarios.” When the present resembles the past, we can calculate the probability of future outcomes by analyzing past historical data. But, the situation is very different when trying to formulate strategy in the face of high uncertainty because when situations lack analogies to the past, we have trouble envisioning how the future will play out.

The object of scenario planning is to assemble all known facts, trends, and other information and use them to create a number of alternative stories rich in detail, typically three or four, about plausible future outcomes. Military strategists first developed scenario planning in the 1950s to help them envision the impact of a nuclear war between the US and the Soviet Union, - something that was clearly way beyond anyone’s experience.

Scenario planning has continued to be used by military and government policy strategists. For example, every four years since 1997, the US National Intelligence Council has been publishing a Global Trends report outlining the key trends that will shape the world over the following twenty years. The reports illustrate how the future might play out by developing different scenarios, typically three or four, each embodying the key issues, trends, decisions and uncertainties that will shape the global landscape twenty years into the future.

In the business world, scenario planning has become a valuable tool for strategic thinking over the mid- to long-term planning horizon. Alternative technology, market and political environments can be described in a few discrete scenarios while organizing the richness and range of possibilities into plausible future outcomes.

The Future of Self-Driving Cars and Robotaxis

MIT professor emeritus Rodney Brooks has been posting an annual Predictions Scorecard  in rodneybrooks.com since January 1, 2018, where he predicts future milestones in three technology areas: AI and robotics, self driving cars, and human space travel. He also reviews the actual progress in each of these areas to see how his past predictions have held up. On January 1 he posted his 2025 Predictions Scorecard.

Two weeks ago I discussed Brooks’ predictions on “The Future of AI and Robotics.” I now want to discuss his predictions on The Future of Self-driving Cars and Robotaxis.

Fully autonomous vehicle (AVs) may well be the quintessential symbol of our AI/robotics age. Cars are a major part of our daily lives. Unlike other AI and robotics applications, a self-driven vehicle is a concept that requires little explanation, — something we can all quickly grasp. It wasn’t that long ago that the notion of an AV driving us around while we read or sleep would have felt like the stuff of science fiction. This is no longer the case.

Brooks wrote that “the definition, or common understanding, of what self driving cars really means has changed since my post on predictions seven years ago. At that time self driving cars meant that the cars would drive themselves to wherever they were told to go with no further human control inputs. Now self driving cars means that there is no one in the drivers seat, but there may well be, as is the case in all cases so far deployed, humans monitoring those cars from a remote location, and occasionally sending control inputs to the cars.”

A Roadmap for Long Term Resilience and Innovation in Global Supply Chain Networks

“In the complex and rapidly evolving business landscape and volatile global market conditions, supply chains cannot thrive in isolation,” begins “Building Constellations of Value — Transforming Supply Chains to Drive Collaboration, Innovation, and Long-Term Value,” a recently published report by the Digital Supply Chain Institute (DSCI), an organization focused on the evolution of enterprise supply chains. “Companies are increasingly recognizing the need for collaborative networks — Constellations of Value (CoV) — that bring together diverse stakeholders across a supply chain to tackle shared challenges and unlock new growth opportunities.”

The CoV framework is based on inputs from DSCI members and other global supply chain leaders. “A key insight from this effort was that successful supply chains are no longer defined solely by their internal efficiencies; instead, they must be designed as collaborative networks that drive shared value, resilience, and innovation.” 

I couldn’t agree more.

As I’ve previously written, I first became interested in global supply chains in the 1990s, when the explosive growth of the Internet was driving a new golden age of globalization. In his 2005 international best-seller The World is Flat, Thomas Friedman explained that the rise of global supply chains was one of the key forces that contributed to flattening the world, along with the Internet, open source software, and related technologies.

The Future of AI and Robotics — 2025 Predictions Scorecard

On January 1, MIT professor emeritus Rodney Brooks published his 2025 Predictions Scorecard. A member of  MIT’s faculty since 1984, Brooks was director of the MIT AI Lab from 1997 to 2003, and the founding director of MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) from 2003 until 2007. He’s also been a robotics entrepreneur, having started a number of companies, including iRobot, Rethink Robotics, and Robust.AI.

Brooks has posted a Predictions Scorecard every year since 2018, where he makes predictions about  future milestones in three technology areas that he closely follows: robotics, AI, and machines learning; self driving cars; and human space travel. He also reviews the actual progress in each of these areas to see how his past predictions have held up, and promises “to review them at the start of the year every year until 2050 (right after my 95th birthday), thirty two years in total” in order to hold himself accountable for those predictions. “How right or wrong was I?”

In his 2023 Predictions Scorecard, Brooks explained that he makes his predictions because he’s seen “an immense amount of hype about these three topics, and the general press and public drawing conclusions about all sorts of things they feared (e.g., truck driving jobs about to disappear, all manual labor of humans about to disappear) or desired (e.g., safe roads about to come into existence, a safe haven for humans on Mars about to start developing) being imminent. My predictions, with dates attached to them, were meant to slow down those expectations, and inject some reality into what I saw as irrational exuberance.”

Let me summarize what he wrote in his 2025 Predictions Scorecard in the area I most closely follow:

What Do We Know About the Economics of AI?

A few weeks ago, MIT News published an article on “What do we know about the economics of AI?” The article is based on the research of MIT Nobel laureate professor Daron Acemoglu, who in October was awarded the 2024 Nobel Memorial Prize in Economic Sciences along with his MIT colleague Simon Johnson and University of Chicago economist James Robinson. “For all the talk about artificial intelligence upending the world, its economic effects remain uncertain,” said the article. “There is massive investment in AI but little clarity about what it will produce.”

Professor Acemoglu has long been conducting research on the impact of technology on the economy. “Artificial intelligence (AI) has captured imaginations. Promises of rapid, even unparalleled, productivity growth as well as new pathways for complementing humans have become commonplace,” he wrote in “The Simple Macroeconomics of AI,” a research paper published in May of 2024.

“AI will have implications for the macroeconomy, productivity, wages and inequality, but all of them are very hard to predict. This has not stopped a series of forecasts over the last year, often centering on the productivity gains that AI will trigger. Some experts believe that truly transformative implications, including artificial general intelligence (AGI) enabling AI to perform essentially all human tasks, could be around the corner. Other forecasters are more grounded, but still predict big effects on output.”

The World Ahead in 2025

A few weeks ago, The Economist published “The World Ahead 2025,” its 39th annual issue that examines the trends and events that will likely shape the coming year. Last year, the overriding message from “The World Ahead 2024” was the mind-boggling speed at which life was changing, from the Middle East turmoil and economic uncertainty to the redrawing of the global energy-resources map and the rapid progress in artificial intelligence (AI).

“It seems appropriate that 2025 has been designated the year of quantum science and technology by the United Nations,” wrote World Ahead editor Tom Sandage  in his introduction to this year’s issue. “Because like Schrödinger’s cat, which (in a quantum thought-experiment) was both alive and dead at the same time inside a closed box, 2025 has hovered in a superposition of two very different states, defined by the outcome of America’s election. Now the ballot boxes have been opened, the world knows which 2025 to expect: the one where Donald Trump returns to the White House. With that uncertainty resolved, here are ten themes to watch in the coming year.”

What Will it Take to Get AI Right? A 2050 Thought Experiment

“During this period of astonishing technical progress and public engagement on artificial intelligence, the world has been grappling with how to get AI right — and for good reason,” wrote James Manyika, in “Getting AI Right: A 2050 Thought Experiment.” Manyika is Senior Vice President of Research, Technology and Society at Google and Chairman Emeritus of the McKinsey Global Institute. His essay is part of The Digitalist Papers, a series of 12 essays on the future of the age of AI.

“[T]he world has been grappling with how to get AI right—and for good reason,” noted Manyika. “AI is a foundational and general-purpose technology with the potential to benefit people and society in a range of ways”:

• assist people with everyday tasks, help them access and benefit from the world’s information and knowledge, and pursue their most ambitious, productive, and creative endeavors;
• contribute to economic progress by enabling entrepreneurs, powering small and large businesses, and fueling innovation and productivity;
• accelerate scientific advances in fields ranging from medicine to materials, physics, climate sciences and more; and
• help make progress on many of humanity’s most pressing challenges and opportunities, from food security to health and well-being.

“At the same time there are concerns about AI’s development, deployment, and use, including: robustness, accuracy, bias, and privacy risks; risks from misapplication and misuse; potentially complex impacts on society from jobs to education and democracy, and the possibility of unintended or unforeseen consequences; and challenges of alignment with human preferences and human flourishing as AI becomes more capable. Such concerns, if unaddressed, could create information hazards, lead to safety and security risks, and cause harm.”

The Evolution of Global Supply Chains

“The central purpose of supply chains is to fulfill the needs and wants of humanity — delivering all the food, medicine, energy, apparel, and other worldly goods needed by the eight billion people in the planet,” wrote MIT professor Yossi Sheffi in the preface of his 2023 book, The Magic Conveyor Belt: Supply Chains, A.I. and the Future of Work. “The societal and economic spasms of the early 2020s highlighted the crucial role of world-spanning supply chains in the modern global economy, as well as the growing role of digital technology, including A.I. and automation, in the future economy.”

Lately, I’ve been thinking a lot about supply chains. Not only are global supply chains among the most complex of man-made systems, but supply chains are nowhere near ready to be transformed because many of the tasks and processes involved in supply chains are still in the early stages of digitalization. As I think about my personal experience in the evolution of the IT industry, the successful collaboration around standards spurred a new era of productivity and growth. We have yet to see a similar catalyst in the supply chain ecosystem.

I’ve been following the evolution of global supply chains for the past three decades. The 1990s ushered in a golden age of globalization, when the world seemed to be coming together. Thomas Friedman’s The World is Flat became an international best-seller in 2005, nicely explaining what globalization was all about, including the key forces that contributed to flattening the world, from the explosive growth of the Internet to the rise of just-in-time global supply chains. The very nature of the global organization was transformed, with IBM CEO Sam Palmisano declaring in a 2006 Foreign Affairs article that “The Globally Integrated Enterprise” was destined to become the corporate model of the future.

My interest in global supply chains increased significantly in the mid 2010s, fueled by my fascination with enterprise blockchain technologies. In 2016, the World Economic Forum (WEF) included blockchain in its annual list of Top Emerging Technologies. The WEF report compared blockchain to the Internet, noting that “Like the Internet, the blockchain is an open, global infrastructure upon which other technologies and applications can be built. And like the Internet, it allows people to bypass traditional intermediaries in their dealings with each other, thereby lowering or even eliminating transaction costs.”

The Breakthroughs AI Needs

A recent issue of The Economist featured a technology focus on “The breakthrough AI needs” with eight articles on the topic. “Two years after ChatGPT took the world by storm, generative artificial intelligence seems to have hit a roadblock,” said the issue’s lead article. “The energy costs of building and using bigger models are spiralling, and breakthroughs are getting harder. Fortunately, researchers and entrepreneurs are racing for ways around the constraints. Their ingenuity will not just transform AI. It will determine which firms prevail, whether investors win, and which country holds sway over the technology.”

This is frightening for investors who’ve bet big on AI, but there is no reason to panic, said the article. “Plenty of other technologies have faced limits and gone on to prosper thanks to human ingenuity,” it added. “Already, developments in AI are showing how constraints can stimulate creativity.” In particular, The Economist mentions two such major innovations: the development of chips with the special purpose architectures needed to train and run AI models as fast and energy efficient as possible, and the development of smaller, more specialized domain specific models that consume a lot less energy than the very large models that rely on brute force computational power.

Let me discuss each of these two innovations.

An Open Architecture for Health Data Interoperability

I’ve been a member of the Linux Foundation (LF) Research Advisory Board for the past three years. At a September, 2023 meeting of the Advisory Board, we discussed that one of the few industry sectors in which the LF doesn’t have projects is healthcare, — despite the size and importance of healthcare around the world. The LF had launched open source healthcare initiatives in the past, but they all eventually failed.

A paper on Digital Health in the 21st Century published in June of 2022 by the US National Academy of Medicine noted that “Revolutionary advances in digital health are transforming health, medicine, and biomedical science, and redefining and re-engineering the tools needed to create a healthier future,” including artificial intelligence, cloud computing, and the increased adoption of electronic health records (EHRs). 

So, perhaps it’s time to try again, and LF Research made the decision to launch a healthcare study, led by Anna Hermansen, Researcher and Ecosystem manager in LF Research, to understand the obstacles for open source adoption and interoperable data standards in the healthcare sector. A year later, in October of 2024, Hermansen released the study’s findings and recommendations in a report, “An Open Architecture for Health Data Interoperability: How Open Source Can Help the Healthcare Sector Overcome the Information Dark Ages.”

Who Will Be Proved Right? The Bullish or the Skeptic AI Camp?

A few weeks ago I posted “AI Advances Aren’t Likely to Occur Nearly as Quickly as Many Believe,” a blog focused on a research paper, “The Simple Macroeconomics of AI,” by MIT professor Daron Acemoglu, — a co-recipient of the 2024 Nobel Memorial Prize in Economic Sciences. The blog referenced a  NY Times article, “What if the A.I. Boosters Are Wrong?,” which discussed Acemoglu’s skeptical views that AI will supercharge economic productivity by the end of this decade, as well as “Gen AI: Too Much Spend, Too Little Benefit?, a report by Goldman Sachs Research senior strategist Allison Nathan, which included a number of insightful interviews, including one with Professor Acemoglu.

As I wrote in the blog, Nathan started her interview by asking why his paper argued that the upside to US productivity and GDP growth from GenAI will likely prove more limited that those of other predictions, including some from Goldman market analysts. “Why are you less optimistic on AI’s potential economic impacts?,” she asked him.

Acemoglu replied that the forecast differences are primarily about the timing of AI’s economic impacts rather than about the ultimate promise of the technology. “Generative AI has the potential to fundamentally change the process of scientific discovery, research and development, innovation, new product and material testing, etc. as well as create new products and platforms.” But the economic impact of historically transformative technologies like AI take time to play out.

I’d like to now discuss another recent NYT article, “Will A.I. Be a Bust?,” which was focused on the views of Jim Covello, Goldman Sachs’ head of stock research, one of Wall Street’s leading AI skeptics. “Three months ago, he jolted markets with a research paper that challenged whether businesses would see a sufficient return on what by some estimates could be $1 trillion in A.I. spending in the coming years,” said the article.

The 2024 State of Open Source in Financial Services

“For the financial services industry, open source collaboration is a strategically relevant practice” said “The 2024 State of Open Source in Financial Services,” the fourth annual research report by the Fintech Open Source Foundation (FINOS) in partnership with Linux Foundation Research. Over the past four years, this study has aimed to provide concrete industry-wide trends about the state of open source adoption in the financial services industry. “Even in the relatively short time since studying sector trends in open source began, we’ve seen encouraging signals for the financial services industry that there are numerous benefits to realize through increased collaboration at the precompetitive level of technology stacks and that more participants in the sector are coming on board.”

As in prior years, the report is based on empirical evidence, including a global survey, industry data, and insights from interviews with senior finance and IT leaders.

The global survey was conducted in May and June of 2024 and received responses from 249 qualified individuals, — 39% from North America, 28% from Europe, and 29% from Asia-Pacific countries. 58% of respondents worked in financial institutions, and 35% were employed by providers of financial and IT services.

“Altogether, the data and the trends collectively underscore the growing importance of open source in driving innovation, enhancing efficiency, and fostering collaboration. Our findings reveal that financial services organizations are not only consuming open source software at increased levels to achieve a variety of objectives but also contributing back to the community in meaningful ways … For stakeholders from capital markets, retail or private banking, asset management, insurance, or for those fintech leaders delivering innovative solutions for the sector, there’s more reason than ever to participate in open source ecosystems.”

The Top 10 Emerging Technologies of 2024

Since 2011, the World Economic Forum (WEF) has published an annual list of the Top Ten Emerging Technologies that are likely to be disruptive within three to five years. The Top 10 Emerging Technologies of 2024 was released by the WEF in June. The report is based on the insights of over 300 world-leading academics and researchers from the WEF’s Network of Global Future Councils and University & Research Network, and from the Frontiers network of over 2,000 chief editors from top global institutions.

A survey was distributed to members of these various networks asking them to propose technologies for the 2024 Top 10 list. The survey received over 300 valid technology nominations from 29 countries. After an initial screening, the Top 10 Emerging Technologies Steering Group was presented with a curated list of 70 technologies from which the final ten were selected based on four criteria:

• Novelty: The technology is emerging and at an early stage of development but is not yet widely used.
• Applicability: The technology is potentially of significant use and benefit to societies and economies.
• Depth: The technology is attracting increased development, investment, and excitement.
• Power: The technology is potentially game-changing to established ways and industries.

AI Advances Aren’t Likely to Occur Nearly as Quickly as Many Believe

“What if the A.I. Boosters Are Wrong?,” asked a recent NY Times article in its title. “A skeptical paper by Daron Acemoglu, a labor economist at M.I.T., has triggered a heated debate over whether artificial intelligence will supercharge productivity,” it added. “The bullish camp has great  hopes for AI,” such as whether the technology will usher the next industrial revolution which could help wipe out poverty around the world. “But if the boosters are wrong, it could be trouble for the developed world, which is in desperate need of a productivity breakthrough as its work force ages and declines.”

(Before continuing with the blog, let me mention that on October 14, 2024, Professor Acemoglu was awarded the 2024 Nobel Memorial Prize in Economic Sciences along with his MIT colleague Simon Johnson, and University of Chicago economist James Robinson.)

“Artificial intelligence (AI) has captured imaginations,” wrote Acemoglu in “The Simple Macroeconomics of AI,” the skeptical paper referenced in the NYT article. “Promises of rapid, even unparalleled, productivity growth as well as new pathways for complementing humans have become commonplace. There is no doubt that recent developments in generative AI and large language models that produce text, information and images — and Shakespearean sonnets — in response to simple user prompts are impressive and even spellbinding.”

“AI will have implications for the macroeconomy, productivity, wages and inequality, but all of them are very hard to predict,” he added. “This has not stopped a series of forecasts over the last year, often centering on the productivity gains that AI will trigger. Some experts believe that truly transformative implications, including artificial general intelligence (AGI) enabling AI to perform essentially all human tasks, could be around the corner.” But, according to Acemoglu, AI will contribute only modest improvements to worker productivity and will add no more than 1 percent to U.S. economic output over the next decade.

Large Language Models (LLMs): Mistaking Engineering Achievements for Human Linguistic Competence

“Mistaking the impressive engineering achievements of LLMs for the mastering of human language, language understanding, and linguistic acts has dire implications for various forms of social participation, human agency, justice and policies surrounding them,” wrote cognitive scientists Abeba Birhane and Marek McGann in a recent paper, “Large models of what? Mistaking engineering achievements for human linguistic agency.” “Hyperbolic claims surrounding LLMs often (mis)uses terms that are naturally applied to the experiences, capabilities, and characteristics of human beings.”

LLMs impressive abilities to respond to, and generate natural language are based on their training using massive language data sets generally sourced from the World Wide Web. The training involves breaking down text or speech into tokens, typically a few characters in length, to develop a statistical model of language. Powerful statistical techniques and lots of computational power are then used to analyze the relationship between billions of tokens in order to generate grammatically valid sequences of token concatenations in response to a question or prompt.

Birhane and McGann note that “the processing of datasets and the generation of output are engineering problems, word prediction or sequence extension grounded in the underlying distribution of previously processed text. The generated text need not necessarily adhere to ‘facts’ in the real world,” which is why LLMs are prone to hallucinations, a response generated by its algorithms that, while statistically and grammatically correct, is actually false or misleading real-world information.

Will AI Devour Software Engineering (SE)?

“Social media provide a steady diet of dire warnings that artificial intelligence (AI) will make software engineering (SE) irrelevant or obsolete,” wrote CMU computer scientists Eunsuk Kang and Mary Shaw in their September, 2024 paper, “tl;dr: Chill, y’all: AI Will Not Devour SE.” [tl;dr: a summary of a longer text]. “To the contrary, the engineering discipline of software is rich and robust; it encompasses the full scope of software design, development, deployment, and practical use; and it has regularly assimilated radical new offerings from AI.”

“Current AI innovations such as machine learning, large language models (LLMs) and generative AI will offer new opportunities to extend the models and methods of SE. They may automate some routine development processes, and they will bring new kinds of components and architectures. If we’re fortunate they may force SE to rethink what we mean by correctness and reliability. They will not, however, render SE irrelevant.”

“What’s with the cries of alarm about AI’s threats to SE?,” asked Kang and Shaw. Social media is full of “angst about the imminent demise of SE at the hands of AI.” Their paper referenced a few recent articles about AI’s potential threat to SE: “The End of Programming,”which noted that “The end of classical computer science is coming, and most of us are dinosaurs waiting for the meteor to hit”; “ChatGPT Will Replace Programmers Within 10 Years”; and “NVIDIA CEO says the future of coding as a career might already be dead in the water with the imminent prevalence of AI.”

“It seems that either AI systems are so different from ‘regular’ software systems that SE knowledge has become obsolete or irrelevant, or else AI will soon take over programming, and by extension software development.” I found this last comment very telling, because, as I read related articles, — like “The rise — and fall — of the software developer” and “Tech Jobs Have Dried Up—and Aren’t Coming Back Soon,” it became clear that software engineering is not a well understood discipline and is often confused and used interchangeably with computer programming and software development, important tasks whose scope is significantly more limited than software engineering.

Digital Twins Are Changing the World

“Digital twins are fast becoming part of everyday life,” said the lead article in the August 31 issue of The Economist. A Digital Twin is essentially a computerized companion to a real-world entity, be it an industrial physical asset like a jet engine, an individual’s health profile, or a highly complex system like a city.

“Digital twins began as basic computer models of physical objects and systems,” explained The Economist. “As computers have become more powerful, twins have become more sophisticated. Complex design and modelling software means many physical objects initially take shape in the virtual world. Small sensors, capable of measuring all sorts of things, feed twins with real-time data, ensuring that they mirror their physical counterparts.”

Digital Twins help bring the physical and digital worlds closer to each other. Digital technologies, internet connectivity, and other innovations are designed right into the products, creating a new era of smart connected products. Massive amounts of product-usage data can now be gathered, stored and analyzed by applications in the product’s cloud counterpart. The use of AI takes all this much further, allowing virtual models to both simulate and optimize activities in the real world.

The Three Laws of Robotics

Earlier this year, MIT professor emeritus Rodney Brooks gave the closing keynote at the 2024 MIT Sloan CIO Symposium. Professor Brooks was director of the MIT AI Lab from 1997 to 2003, and was the founding director of MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) from 2003 until 2007. A robotics entrepreneur, he’s founded of a number of companies, including iRobot, Rethink Robotics, and Robust.AI.

Given that AI was the overall theme of this year’s MIT Symposium, Brooks focused his keynote on “What Works and Doesn't Work with AI.” He reminded the audience that “AI has been on the verge of changing everything” ever since the field was founded in the 1950s, and explained why so many AI predictions have turned out so wrong.

I wrote about his excellent presentation in “Artificial Intelligence: Realistic Expectations vs. Irrational Exuberance,” and concluded the blog with what Brooks called “My Three Laws of Artificial Intelligence”:

• When an AI system performs a task, human observers immediately estimate its general competence in areas that seem related. Usually that estimate is wildly overinflated.
• Most successful AI deployments have a human somewhere in the loop (perhaps the person they are helping) and their intelligence smooths the edges.
• Without carefully boxing in how an AI system is deployed there is always a long tail of special cases that take decades to discover and fix. Paradoxically all those fixes are AI-complete themselves.

At the end of July, Brooks posted two new essays in his website. One was “Rodney Brooks’ Three Laws of Artificial Intelligence,” where he elaborated on his Three Laws of AI. I’d like to now discuss his second essay, “Rodney Brooks’ Three Laws of Robotics.”

Is the AI Revolution Already Losing Steam?

“It sure seems like the AI hype train is just leaving the station, and we should all hop aboard,” said “The AI Revolution Is Already Losing Steam,” a recent WSJ article by technology columnist Christopher Mims. “But significant disappointment may be on the horizon, both in terms of what AI can do, and the returns it will generate for investors. The rate of improvement for AIs is slowing, and there appear to be fewer applications than originally imagined for even the most capable of them. It is wildly expensive to build and run AI. New, competing AI models are popping up constantly, but it takes a long time for them to have a meaningful impact on how most people actually work.”

As we’ve learned since the advent of the Industrial Revolution, there’s generally been a significant time lag between the broad acceptance of a major new transformative technology and its ensuing impact on industries and economies. After transitioning from the R&D labs to the marketplace, it takes considerable time, — often decades, — for the products, services, and business models based on the new technologies to be widely embraced across economies and societies, and for their full benefits to be realized.

The reason, explained economists  Erik Brynjolfsson, Daniel Rock, and Chad Syverson in “The Productivity J-Curve,” a 2018 NBER working paper, is that while historically transformative technologies have the potential to radically change the economic environment, “realizing that potential requires larger intangible and often unmeasured investments and a fundamental rethinking of the organization of production itself.”

Their paper identifies two phases, investment and harvesting, in the life cycle of a historically transformative technology. In general, these technologies require massive complementary investments, such as business process redesign, co-invention of new products and business models, and the re-skilling of the workforce. Moreover, the more transformative the technologies, the longer it takes for them to reach the harvesting phase when they are widely embraced by companies and industries across the economy. Such an evolution has been has been pithily captured in what’s become known as Amara’s Law: We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run.

The Geek Way: A Better Way to Run a 21st Century Company

“Have you recently heard the CEO or other senior executive of a company say something like, we’re a technology company now?” asked Andrew McAfee ealier this year in his keynote at the 2024 MIT IDE Annual Conference. McAfee is a Principal Research Scientist at the MIT Sloan School of Management, and co-founder and co-director of MIT’s Initiative on the Digital Economy. “And in particular, when you heard that person say that, did you think to yourself that’s really not a technology company at all?”

He explained that what CEOs are trying to convey when they say that they run a technology company is that their company needs to be managed very differently now. The pressure to innovate, remain agile and execute at the very high level that technology companies have felt for the past two decades has actually been getting more intense and now applies to their own companies, regardless of the industry they are in.

McAfee’s keynote is based on his November, 2023 best seller The Geek Way: The Radical Mindset That Drives Extraordinary Results. The book’s thesis is succinctly articulated in its introduction: “A bunch of geeks have figured out a better way to run a company. As a result, they’re taking over the economy. And they’re just getting started.”

The 2024 State of Talent in the IT Sector

A few months ago the Linux Foundation (LF) published the “2024 State of Tech Talent Report,” a comprehensive analysis of talent management practices in the IT sector. The study is based on a web survey conducted by the LF and its partners earlier this year. The survey reached out to individuals responsible for hiring and training IT professionals around the world to better understand the current state of talent acquisition, retention, and management.

The survey analyzed data from 418 respondents carefully chosen to eliminates sample bias and ensure high data quality. Respondents came from companies of different sizes across a variety of vertical industries: 17% worked in organizations with less that 250 employees, 57% worked in organizations from 250 to 4,999 employees, and 26% in organizations with more than 5,000 employees.

 Respondents included IT vendors, service provides, nonprofits, academic institutions, and government organizations from around the world: 30% came from organizations headquartered in the US and Canada, 19% from Japan, 18% from Western Europe, 8% from India, 8% from other Asian countries, 4% from Eastern Europe, and 4% from Latin America.

The Foundation Model Transparency Index

“Foundation models are the epicenter of artificial intelligence (AI) as AI begins to shape how the economy and society function,” begins “The Foundation Model Transparency Index v1.1 (FMTI v1.1),” a May, 2024 paper by a group of researchers from Stanford, MIT, and Princeton under the auspices of Stanford’s Center for Research on Foundation Models (CRFM). “For such a high-impact technology, transparency is vital to facilitate accountability, competition, and collective understanding. As an illustrative example, the current lack of transparency regarding the data used to build foundation models makes it difficult to assess what copyrighted information is used to train foundation models. Governments around the world are intervening to increase transparency.”

In October of 2023, CRFM introduced the Foundation Model Transparency Index (FMTI), a scoring system designed by a multidisciplinary research team to score the transparency of foundation models in order to promote responsible business practices and greater public accountability. The FMTI features three top-level domains, which in turn aggregate 23 subdomains and 100 transparency indicators:

• Upstream domain — the 32 indicators and 6 subdomains used to build the model, including data size and composition, human labor used, training methods, hardware and energy used, and environmental impacts;
  
• Model domain — the 33 indicators and 8 subdomains that identify the properties and function of the foundation model itself, including size, capabilities, limitations, risks, and trustworthiness; and
  
• Downstream domain — the 35 indicators and 9 subdomains that identify how the model is used, including release process, distribution channels, user interface, data protection, feedback mechanisms, and documentation.
  

Regulation by Design: Embedding Regulatory Objectives into the Design of AI Systems

In its early decades, IT helped companies become significantly more productive by automating many of their processes, reengineering their overall operations, and reaching out beyond the boundaries of the enterprise. But, their actual products were largely unaffected. This has all been changing in the 21st century with the rise of the Internet of Things (IoT), Big Data, and Artificial Intelligence, whose combined impact is driving a new era of smart connected products.

As the world’s digital and physical infrastructures converge, digital technologies are designed right into complex products, — e.g., jet engines, power generators, medical equipment, energy pipelines. Massive amounts of usage data can now be gathered over the internet, stored and analyzed by sophisticated applications to help monitor the product and anticipate potential problems. This is particularly important in industries where safety by design can have a significant economic impact and in some cases, literally save lives. Safety by design features are now in widespread use in all kinds of mechanical, electronic, and other complex physical systems.

What if we could similarly embed regulatory objectives directly into the technical design of AI systems?, asked Robert Mahari in  his keynote on Regulation by Design at the recent 2024 MIT IDE Annual Conference. Mahari is pursuing a joint JD-PhD degree. He received a JD at the Harvard Law School in 2022 and is now a PhD candidate in the MIT Media Lab research group led by professor Alex ‘Sandy’ Pentland.

“Compliance and regulation by design represent a risk-management paradigm that’s uniquely suited for AI,” Mahari said. “Intelligent technology design can proactively prevent failures and risks.”

AI systems are the result of a very complex supply chain, including the choice of the large amounts of data needed to train AI models, and the difficulty of testing, predicting, and explaining the models’ expected behavior. As a result, compliance with regulatory and ethical objectives is very hard for the users of AI-based applications who have little control or understanding of how the AI system works. Instead, as is the case with advanced engineering systems, can an AI system monitor how it’s used, identify high risk sessions, alert its developers and overseers, and thus ensure compliance?

Language is a Tool for Communication, Not for Thinking

I often look to biology and evolution for inspiration when thinking about highly complex systems. So I was very intrigued when several weeks ago I learned about the work being done by Elemental Cognition, a company founded in 2015 by my former IBM colleague David Ferrucci. Ferrucci led the team that developed the Watson computer system, which in 2011 won the television quiz show Jeopardy! against two of the world’s top Jeopardy champions. Let me explain.

Large language models (LLMs), chatbots, and generative AI have introduced technologists like me to the fascinating world of human language and cognition. Last year, for example, I learned the difference between four important linguistic concepts in a 2020 research paper by linguistic professors Emily Bender and Alexander Koller.

• Form is any observable expression of language, whether written, spoken, or signed.
• Communicative intent is the purpose a speaker intends to achieve through language, such as to convey some information or just to socialize.
• Meaning is the relation between the form in which the language is expressed and the communicative intent it’s being used to evoke in the listener or reader.
• Understanding is the listener’s ability to capture the meaning that the speaker intends to convey.

The authors wrote that while the success of LLMs and chatbots on many natural language tasks is very exciting, “these successes sometimes lead to hype in which these models are being described as understanding language or capturing meaning.” However, “a system trained only on form has apriori no way to learn meaning.” Research on language acquisition has found that “human language learning is not only grounded in the physical world around us, but also in interaction with other people in that world. … Human children do not learn meaning from form alone and we should not expect machines to do so either.”

How Does the Linux Foundation Work and What Does It Do?

I have long been involved with the Linux Foundation (LF). In 2000, a consortium of companies, including IBM, HP, and Intel, founded the Open Source Development Labs (OSDL), a non-profit organization to support the continued development of Linux and promote its use in enterprise computing. A few years later, in 2007, the OSDL merged with the Free Standards Group to become the Linux Foundation. At the time, I was leading the Linux initiative that IBM had launched in January of 2000 to embrace Linux across all the company’s products and services. In 2011, when the LF marked the 20th anniversary of Linux at its annual Linux conference, I had the privilege of giving a keynote presentation where I recounted my personal involvement with Linux and open source.

Over the past decade, the LF has undergone  a major expansion, as open source became widely accepted as a paradigm for innovation across a full suite of technologies, proving its utility beyond the collaborative development of the Linux operating system. Its impact has been felt across just about all industry sectors in our fast growing digital economy. Consequently, its annual conferences were rebranded Open Source Summits to be more representative of the LF’s expanded open source mission beyond Linux. In April of 2021, the LF announced the formation of Linux Foundation Research to broaden the understanding of open source projects and to explore new opportunities for open source initiatives. Later that year, Hilary Carter, LF’s then VP of Research, invited me to become a member of the LF Research Advisory Board.

I am truly impressed by the current scope of the Linux Foundation. The LF now includes over 1,400 company members, supports over 1000 open source projects in a variety of vertical and horizontal industry sectors, holds over 70 events around the world every year, and provides training and certification programs to equip developers with essential skills, not just for coding, but to support the management and best practices the projects themselves. The mission of the LF was quite easy to explain when it was mostly focused on Linux, but it’s quite a bit more complicated now.

LF’s Executive Director Jim Zemlin recently addressed the difficulty of explaining what the LF does and how it works in his June 13 weekly newsletter to the LF board and team members. Zemlin wrote that beyond the technical community, many don’t understand much about how the LF operates, such as the important role of the LF’s company members in sponsoring open source projects, and that generally, the key maintainers of LF’s largest projects are not volunteers but full-time employees of member companies.

The Emerging Data-Centric AI Era: Where Will All the Data Come From?

The data-centric AI era started about 25 years ago when the explosive growth of the internet led to what’s become known as big data, that is, the availability of huge amounts of digital data, including text, voice, and images. The 2010s saw major innovations in multi-layered deep learning algorithms, followed in the past few years by the advent of foundation models, including generative AI, large language models, and chatbots.

Given the explosive growth of interest in AI, the demand for data keeps increasing. No matter what an organization hopes to achieve, success is impossible without ready access to high-quality data. But, as data becomes increasingly critical for business success, how are enterprises adjusting to the increasing importance of data? Where is all the data needed to train AI models going to come from? And what are the potential legal and ethical issues that companies need to watch out for?

To explore these important questions, I moderated a panel on The Data Provenance Dilemma at the 2024 MIT Sloan CIO Symposium on May 14. “Generative AI models are highly dependent on the quality and quantity of data used in their training. But, many of these models are being trained on vast, diverse, and inconsistently documented datasets that have been raising serious concerns about the legal and ethical risks involved,” notes the brief for the panel in the Symposium agenda. “The panel will discuss the complexities of using large, diverse datasets in the training of generative AI models, the difficulties involved in understanding the provenance of the data being used, and the kinds of tools and standards needed for the responsible use of these powerful models.”

The panel included Mike Mason, chief AI officer of the technology consultancy Thoughtworks; Shayne Longpre, a doctorate candidate at the MIT Media Lab; and Robert Mahari, who’s pursuing a joint JD-PhD degree at the Harvard Law School and MIT’s Media Lab. Let me summarize the key points we discussed in the panel.

Advancing Global Trade through Open-Source, Digital Bill of Lading

“Imagine you’re a cargo owner in the year 1450,” begins “The multi-billion-dollar paper jam: Unlocking trade by digitalizing documentation,” a McKinsey article published in October of 2022. “You hand over your goods to the ship that will carry them across the world, and are presented with a bill of lading — a piece of paper stating what you’re shipping, where it comes from, and where it’s heading. Fast forward to the year 2022: The world has changed dramatically, but the bill of lading remains relatively unchanged. Today, the bill of lading process is still reliant on the physical transfer of paper records and applies to roughly 40 percent of all containerized trade transactions.”

“Current trade documentation spans many documents and processes, and is a manual, time-consuming, and resource-intensive process for all stakeholders. Documentation for a single shipment can require up to 50 sheets of paper that are exchanged with up to 30 different stakeholders. The bill of lading, issued by carriers to acknowledge receipt of cargo from the shipper, is one of the most important trade documents required for shipping.”

According to McKinsey, the bill of lading accounts for between 10 and 30 percent of total trade documentation costs; adopting an electronic bill of lading could save $6.5 billion in direct costs and enable between $30 billion and $40 billion in new global trade volume. “While the banking and aviation industries have implemented digital standards enabling automated trade systems, shipping has not matured far beyond where it was in the 1400s.”

In a recent article, “Advancing Global Trade through Open-Source Electronic Bill of Lading,” the Digital Supply Chain Institute (DSCI) explained the role of bill of lading in global supply chains. “The bill of lading (B/L) functions as an important role in global trade, providing a tangible record of cargo ownership, conditions, and delivery obligations,” according to the article. “While deeply entrenched in trade practices, the traditional paper-based bill of lading has its shortcomings: inefficiencies stemming from manual processes have led to delays, errors, and a higher susceptibility to loss and fraud.”

The Critical Value of Data to Business Success in the Emerging AI Era

“Data has become an increasingly critical component to business success,” said “Modernizing Data with Strategic Purpose,” a recently published report by the technology consultancy Thoughtworks and MIT Technology Review Insights. “The ability to make timely and informed decisions, create tailored customer experiences, and identify new opportunities has enabled the rise of today’s leading digital companies. And with the recent explosion of interest in AI, the demand for data has only increased. No matter what an organization hopes to achieve, success is impossible without ready access to high-quality data. Despite advances in technologies, extracting and transforming enterprise data into a usable asset remains a tremendous challenge for most organizations.”

The report is based on a survey conducted in early 2024 of 350 senior data and technology executives from organizations in the US, the UK, Germany, Singapore, and Australia that earn $500 million or more in annual revenue, with almost half (47%) having revenue of $10 billion or more. Eight industries are represented in the survey: financial services, energy, health care and life sciences, manufacturing, public sector, publishing and media, retail, and travel and transport.

The survey sheds light on how enterprises are adjusting to the increasing importance of data in our emerging data-centric AI era. The data-centric AI era started about 25 years ago when the explosive growth of the internet led to what’s become known as big data, that is, the availability of huge amounts of digital data, including text, voice, and images. Digital data has continued to grow exponentially, and is now estimated to have reached around 150 zettabytes (10²¹ bytes). All these data has enabled AI to make great strides in the 2010s with multi-layered deep learning algorithms, and more recently with foundation models, including generative AI (GenAI), large language models, and chatbots.

“As AI pushes data modernization to the top of more organizations’ agendas, it will also become increasingly important to ensure that the data strategy is closely aligned to the broader business strategy and data leaders are able to clearly articulate how data and analytics can help business units achieve their objectives,” notes the report. “This will not only ensure critical buy-in, but also put the organization in the best position to achieve its goals and lay a solid foundation for continuous improvement, evolution, and value creation.”

Artificial Intelligence: Realistic Expectations vs. Irrational Exuberance

The 2024 MIT Sloan CIO Symposium took place on Tuesday, May 14 at the Royal Sonesta, a hotel overlooking the Charles River a short walk from the MIT campus in Cambridge, MA. Not surprisingly, AI was the dominant theme in this year’s Symposium, with a number of keynotes and panels on the topic. In addition, a pre-event program was added on the day before the Symposium, which included a number of more informal roundtable discussions on various aspects of AI, such as legal risks in AI deployment, AI as a driver for productivity, and human’s role in AI-augmented workplaces.

The Symposium’s closing keynote, What Works and Doesn't Work with AI, was delivered by MIT professor emeritus Rodney Brooks. Professor Brooks was director of the MIT AI Lab from 1997 to 2003, and was then the founding director of MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) from 2003 until 2007. A robotics entrepreneur, he’s founded of a number of companies, including iRobot, Rethink Robotics, and Robust.AI.

“There is a lot of AI and robotics hype right now,” said the blurb describing Brooks’ keynote in the Symposium agenda. “Amongst it are a few important innovations that may help many businesses. The trick is to figure out what is close to deployable and what is off in an uncertain distant future. This keynote will talk both about the traps all of us fall in to evaluating AI claims and demonstrations, and what to look for specifically which will get you to a solid understanding of the time frame for if and when technology will be ready to deploy.”

Let me summarize his keynote.

An Open Source, Standards-Based Approach for the Future of Supply Chains

In 2018 I participated in a blockchain panel at a conference of senior supply chain executives representing a variety of industries and companies. I welcomed the opportunity to learn how these supply chain executives felt about the potential impact of blockchain technologies in complex supply chains.

At the time, I was convinced that blockchain technologies would be a major next step in the evolution of the internet. Two years earlier, for example, the World Economic Forum (WEF) had named The Blockchain as one of the Top Ten Emerging Technologies for 2016 in its annual list. The WEF report compared the blockchain to the internet, noting that “Like the Internet, the blockchain is an open, global infrastructure upon which other technologies and applications can be built.  And like the Internet, it allows people to bypass traditional intermediaries in their dealings with each other, thereby lowering or even eliminating transaction costs.”

I also believed that supply chain applications would emerge as the killer app for blockchain technologies, that is, an IT application whose value is simple to explain, its use is fairly intuitive, and the application turns out to be so useful that it helps propel the success of new technologies or products. For example, spreadsheets and word processing played a major role in the 1980s adoption of personal computers, as did e-mail and the World Wide Web in the 1990s mainstream success of the internet.

But, it hasn’t quite worked as expected, either for blockchain technologies or for their applications in supply chains. “Blockchain, the technology underpinning bitcoin and other cryptocurrencies, for years has been viewed by some companies as a way to drive industry-transforming projects, among them the tracking of assets through complex supply chains,” according to “Blockchain Fails to Gain Traction in the Enterprise,” an article published in the Wall Street Journal (WSJ). “So far that hasn’t happened.”

Health and AI

The March 30, 2024 issue of The Economist includes a special  section on Health and AI, with six articles on the topic. “Artificial intelligence (AI) is generating excitement and hyperbole everywhere, but in the field of health care it has the potential to be transformational,” said the lead article, “The AI doctor will see you … eventually.” “In Europe analysts predict that deploying AI could save hundreds of thousands of lives each year; in America, they say, it could also save money, shaving $200bn-360bn from overall annual medical spending, now $4.5trn a year (or 17% of GDP).”

Healthcare is a system of coupled systems, encompassing medical and pharmaceutical research; the delivery of healthcare to patients by a variety of practitioners, including hospitals, physicians, nurses, and pharmacists; and the insurance companies and governments that pay for healthcare. It’s a highly regulated industry, which creates major barriers for the sort of innovative start-ups that have been the engine of transformation in other sectors. In advanced economies, a handful of companies, — e.g., Cerner, Epic, Athena, — support large healthcare providers with proprietary platforms that don’t interoperate with each other, making it difficult for smaller companies and startups to develop innovative AI tools and applications as has been the case in other industries.

There’s considerable hope that AI could help us better deal with the inherent complexity of the healthcare sector. “AI systems can enhance diagnostic accuracy and disease tracking, improve the prediction of patients’ outcomes and suggest better treatments,” notes The Economist. “It can also boost efficiency in hospitals and surgeries by taking on tasks such as medical transcription and monitoring patients, and by streamlining administration. It may already be speeding the time it takes for new drugs to reach clinical trials. New tools, including generative AI, could supercharge these abilities.”

However, “although AI has been used in health care for many years, integration has been slow and the results have often been mediocre.” Among the major reasons for the slow progress are the concerns about accuracy and trust necessary to protect patients’ safety. “Improving accuracy and reducing bias in AI tools requires them to be trained on large data sets that reflect patients’ full diversity.” But, while there’s lots of available health data, the data is highly fragmented making it difficult to use properly.

An Economic Framework for Understanding AI

“Artificial intelligence (AI) systems touch the lives of virtually every American,” said “An Economic Framework for Understanding Artificial Intelligence,” an overview of the economic potential of AI which was recently released by the White House Council of Economic Advisors (CEA) as part of their “2024 Economic Report of the President.” “In recent years, AI systems have advanced rapidly as recent developments in computing, data availability, and machine learning models have simultaneously come together to produce rapid improvements.”

“Still, much remains unknown,” the CEA added. “While AI’s capabilities will depend in part on the technology itself, its effects will be shaped by economic, regulatory, and social pressures. How society deploys this technology and what technology-specific guardrails are implemented will be critical factors in determining both the breadth and magnitude of its effects.”

“An economic framework, combined with a basic understanding of AI technology, allows us to make predictions about when, how, and why AI may be adopted. While such a framework can also tell us what broader effects AI adoption may have, applying economic insights to an evolving and proliferating technology like AI is especially challenging. However, it is also especially valuable, because decisions made at the onset of a new technology have a greater influence on its eventual impact.”

The AI Economic Framework is a rather long and comprehensive document. Let me briefly discuss a few of the its key points.

A Deteroriating Global Risk Outlook

“Last year’s Global Risks Report warned of a world that would not easily rebound from continued shocks,” wrote Saadia Zahidi, managing director of the World Economic Forum (WEF), in the Preface of The Global Risks Report 2024. “As 2024 begins, the 19th edition of the report is set against a backdrop of rapidly accelerating technological change and economic uncertainty, as the world is plagued by a duo of dangerous crises: climate and conflict.”

“Underlying geopolitical tensions combined with the eruption of active hostilities in multiple regions is contributing to an unstable global order characterized by polarizing narratives, eroding trust and insecurity,” she added. “At the same time, countries are grappling with the impacts of record-breaking extreme weather, as climate-change adaptation efforts and resources fall short of the type, scale and intensity of climate-related events already taking place.”

The Global Risks Report 2024 is based on two complementary surveys: the Global Risks Perception Survey (GRPS), the WEF’s source of original risk data, which includes the insights and opinions of over 1,500 leaders from academia, business, government, the international community and civil society; and the Executive Opinion Survey (EOS) of over 11,000 leaders from 113 countries who were asked to identify risks that pose the most severe threat to each of their countries over the next two years. In addition, the report includes insights from over 200 risk specialists.

The WEF report analyzed global risks and presented its findings along three distinct timeframes:

• The current term are the risks most likely to present a material crisis on a global scale in 2024.
• The short term considers global risks which are likely to play out over the next two years.
• The long term considers global risks that are likely to be most severe over the next ten years.

What’s the Value of Open Source Software Based on Actual Usage Data?

“The value of a non-pecuniary (free) product is inherently difficult to assess,” wrote Manuel Hoffmann, Frank Nagle, and Yanuo Zhou in their recent working paper, “The Value of Open Source Software.” “A pervasive example is open source software (OSS), a global public good that plays a vital role in the economy and is foundational for most technology we use today.” The paper reminded us that in 2011, VC Mark Andreessen famously argued that software is eating the world. Then two years ago, VC Joseph Jacks added that open source is eating software FASTER than software is eating the world.

“Other recent studies have come to similar conclusions showing that open source software (OSS) — software whose source code is publicly available for inspection, use, and modification and is often created in a decentralized manner and distributed for free — appears in 96% of codebases, and that some commercial software consists of up to 99.9% freely available OSS,” said the paper. “Although in its early days OSS frequently copied features from existing proprietary software, OSS today includes cutting edge technology in various fields including artificial intelligence (AI), quantum computing, big data, and analytics.”

But, despite the increasing importance of OSS to the economy, its overall impact has been difficult to quantify. The reason is that the value created by a good or service is traditionally measured by multiplying the price p times the quantities sold q, but with OSS, the p is generally zero since open source code is publicly available, and its actual usage q has been very hard to measure.

The Case for Open Healthcare Systems

As we’ve learned since the advent of the Industrial Revolution, major technological advances have the potential to radically transform a whole industry, but realizing that potential generally requires large intangible investments and a fundamental rethinking the industry being transformed. So, it’s not surprising that there’s much hope and excitement that the advent of AI as a major 21st century technological advance might, over time, help to radically transform our highly complex and expensive healthcare industry. 

Healthcare is a system of coupled systems, encompassing medical and pharmaceutical research; the delivery of healthcare to patients by a variety of practitioners, including hospitals, physicians, nurses, and pharmacists; and the insurance companies and governments that pay for healthcare. Healthcare costs have continued to rise across the world, especially in the US. Total healthcare spending in 2023 was approximately $4.7 trillion, accounting for 18% of US GDP, and is projected to grow at an average of 5.5% per year to over $7 trillion in 2031, roughly 20% of US GDP.

The potential evolution of healthcare brings to mind my personal experience in the information technology (IT) industry for the past 50 years.

The Uncertain Impact of AI on Economies and Societies

“Artificial intelligence (AI) promises to boost productivity and growth, but its impact on economies and societies is uncertain, varying by job roles and sectors, with the potential to amplify disparities,” said “GenAI: Artificial Intelligence and the Future of Work,” a recently published report by the International Monetary Fund (IMF). “As a positive productivity shock, AI will expand economies’ production frontiers and will lead to reallocations between labor and capital while triggering potentially profound changes in many jobs and sectors. AI offers unprecedented opportunities for solving complex problems and improving the accuracy of predictions, enhancing decision-making, boosting economic growth, and improving lives. However, precisely because of its vast and flexible applicability in numerous domains, the implications for economies and societies are uncertain.”

“The impact of AI is also likely to differ significantly across countries at different levels of development or with different economic structures,” adds the IMF report. “Advanced economies, with their mature industries and service-driven economies, typically have a higher concentration of jobs in sectors that require complex cognitive tasks. These economies are therefore both more susceptible to, yet better positioned to benefit from, AI innovations.”

“Conversely, emerging market and developing economies, often still reliant on manual labor and traditional industries, may initially face fewer AI-induced disruptions. However, these economies may also miss out on early AI-driven productivity gains, given their lack of infrastructure and a skilled workforce. Over time, the AI divide could exacerbate existing economic disparities, with advanced economies harnessing AI for competitive advantage while emerging market and developing economies grapple with integrating AI into their growth models.”

The Promise of Digital Health in the 21st Century

In 2021 the Linux Foundation (LF) announced the formation of Linux Foundation Research, a new organization to broaden the understanding of open source ecosystems and to explore opportunities for new open source initiatives. Given my long involvement with Linux, the Linux Foundation and open source communities in general, I was invited to become a member of the newly created LF Research Advisory Board (RAB).

At a September, 2023 meeting of the Advisory Board, we discussed that one of the few industry sectors in which the LF does not have projects is healthcare, despite its size and importance around the world. The LF had launched healthcare initiatives in the past, but they all eventually failed. At that meeting, LF Research launched a study to understand what we should be doing in the healthcare sector. I’m involved in the study, and for the past several months, I’ve been learning about the opportunities of digital technologies in healthcare systems by reading a number of articles and talking to key experts.

“The Promise of Digital Health: Then, Now, and the Future,” a discussion paper published in June of 2022 by the US National Academy of Medicine is one of the articles that I found most instructive. The paper is based on the contributions of 25 member of the National Academy of Medicine (NAM), including two of the experts I’ve recently spoken with, John Halamka and Peter Lee.