Irving Wladawsky-Berger

A collection of observations, news and resources on the changing nature of innovation, technology, leadership, and other subjects.

ABOUT

RECENT POSTS

CATEGORIES

Subscribe to this blog via email

  • AI Interpretability: Understanding How AI’s Black-Box Brain Works

    No comments on AI Interpretability: Understanding How AI’s Black-Box Brain Works

    “When Deep Blue, IBM’s chess-playing supercomputer, beat Garry Kasparov in 1997, computers were still just computers,” noted a recent NY Times Magazine article, “We Don’t Really Know How A.I. Works. That’s a Problem,” by freelance writer Oliver Whang. Deep Blue determined the best next move by simulating and assigning values to board positions up to 12 moves ahead — amounting to billions of positions — using algorithms explicitly programmed by its designers. “There was no mystery around what was going on inside them,” wrote Whang, “even though they were, in a way, intelligent.”

    Fifteen years later, everything changed. In 2012, researchers at the University of Toronto developed AlexNet, a neural-network system that identified objects in images far more accurately than previous approaches. AlexNet’s success transformed AI research and accelerated the adoption of deep neural networks across a wide range of applications.

    But there was a catch. Unlike Deep Blue, neural networks operate largely as black boxes. As these models become larger and more capable, they also become increasingly difficult to understand — even for the researchers building them. These systems represent a new kind of machine intelligence whose internal reasoning processes remain poorly understood, even by the researchers building them.

    This has led to the growing field of AI interpretability, whose goal is to better understand how modern AI systems actually work internally, especially as they are increasingly deployed in high-stakes applications ranging from healthcare and finance to law enforcement and military systems. (more…)

    , , , , , , ,

  • The Evolution of Digital Market in the Age of AI Agents

    No comments on The Evolution of Digital Market in the Age of AI Agents

    “AI agents — autonomous systems that perceive, reason, and act on behalf of human principals — are poised to transform digital markets by dramatically reducing transaction costs,” wrote Peyman Shahidi, Gili Rusak, Benjamin Manning, Andrey Fradkin, and John Horton in their recent NBER working paper, “The Coasean Singularity? Demand, Supply, and Market Design with AI Agents.”

    The authors argue that the broad adoption of AI agents will have far-reaching economic consequences. While the exact outcomes remain uncertain, the underlying forces are familiar: supply and demand will shape how agents are deployed, and technological change will alter the relative costs of economic activities. To understand these shifts, they turn to a foundational insight from economics — Ronald Coase’s 1937 theory of the firm.

    Why do firms exist? This question has long been central to economic theory, most notably to the theories of Ronald Coase, the eminent British economist and recipient of the 1991 Nobel Prize in economics. In his 1937 seminal paper, “The Nature of the Firm,” Coase explained that in principle, firms could rely entirely on open markets to procure goods and services efficiently. In practice, however, markets are not frictionless. Transaction costs — searching for suppliers, negotiating contracts, coordinating work, and managing intellectual property — make purely market-based coordination costly and inefficient. Firms emerged as a way to reduce these costs and organize economic activity more effectively.

    A firm will expand as long as it is cheaper to perform additional activities internally than to contract them out in the marketplace. But this expansion has limits. As firms grow, they often become more hierarchical and bureaucratic, which can slow decision-making and reduce adaptability. Successful organizations therefore seek an optimal balance between internal production and external sourcing — the classic “make-or-buy” decision. (more…)

    , , , , , ,

  • The Future We Should Want from the AI Revolution

    No comments on The Future We Should Want from the AI Revolution

    “Transformative AI has revolutionary potential,” wrote Stanford University postdoctoral fellow Gabriel Unger in “Economic Possibilities for Artificial Intelligence,” one of the 21 essays in Volume 2 of The Digitalist Papers — a roadmap of the relentlessly advancing capabilities of the AI revolution.

    Unger adds that the AI revolution has arrived at a particularly significant moment. “America has spent most of the past 50 years in a period of stagnant productivity growth.” Income inequality has risen while trust in institutions — and in each other — has declined. “There is both the economic fact of objectively diminished prospects, and the social fact of rising pessimism in opinion polls about our present and our future.” Our real opportunity, he writes, “is about how we might be able to use a profoundly transformative technology to help rescue ourselves from our decades of increasing dissatisfaction and diminished expectations, in the service of a more promising economic and social life.”

    Unger argues that whether we can fulfill this potential depends critically on answering three broad questions:

    • Can we articulate a compelling shared vision of a future with Transformative AI (TAI) that ordinary people will find exciting and compelling?
    • Can we settle on a theory of economic growth that helps us realize the full economic potential of AI?
    • And can we redesign education and strengthen human connections in the age of Transformative AI?

    “There are plenty of AI optimists and AI pessimists,” notes Unger. The pessimists’ fatalism stems from their belief that the future has already been decided, while the optimists believe that the future remains open. The optimists argue that we should turn our attention “to the unresolved questions most important to a better future from AI.” (more…)

    , , , , , ,

  • Forecasting the Economic Impact of AI

    No comments on Forecasting the Economic Impact of AI

    Over the past few years, I’ve written a number of posts on the potential economic impact of artificial intelligence. One theme that has consistently stood out is the wide divergence of views on how transformative that impact will be. While there is broad agreement that AI capabilities are advancing rapidly, there is far less agreement on what those advances will mean for economic growth, jobs, and productivity.

    “Among tech evangelists in Silicon Valley, it has become conventional wisdom that artificial intelligence will rapidly reshape the labor market, for better or worse. Economists, however, have often discussed A.I.’s impact with a skepticism bordering on dismissiveness,” wrote Ben Casselman, chief New York Times economics correspondent, in a recent article, “Economists Once Dismissed the A.I. Job Threat, but Not Anymore.”

    “Recently, however, the message from economists has undergone a subtle change,” he added. “Most still do not see much evidence that A.I. is disrupting the job market. But they are starting to take seriously the possibility that it could someday soon. If it does, they are worried that policymakers are not ready to respond.”

    The article cites a March 2026 working paper, “Forecasting the Economic Effects of AI,” based on a survey of how AI might affect the U.S. economy over the next 5 and 25 years. The survey, conducted from October 2025 to February 2026, tracked the views of 69 leading economists, 52 AI industry and policy experts, 38 highly accurate superforecasters, and 401 members of the general public.

    The survey results were summarized in a Substack essay, “Forecasting the Economic Effects of AI,” by the Forecasting Research Institute (FRI). (more…)

    , , , , , , ,

  • The Democratization of (Artificial) Intelligence

    No comments on The Democratization of (Artificial) Intelligence

    “New technologies don’t emerge in a vacuum,” wrote Sarah Friar, Chief Financial Officer of OpenAI in “The Democratizing of Intelligence,” one of 21 essays in The Digitalist Papers Volume 2, — a roadmap of the relentlessly advancing capabilities of the AI revolution. “In the beginning, they often reflect and amplify the inequalities of their time — access is shaped by where you live, what you earn, and the systems around you. Then, broader diffusion takes hold. That is what we have seen in the past, with previous waves of technological progress. And that is what we see with AI today.”

    Friar referenced a recent study that tracked the use of ChatGPT over the three years since its launch in November, 2022. As has generally been the case with new technologies, the majority of early adopters were young, highly educated males in living in high-income countries. But by mid-2025, the story had started to change. ChatGPT had already been adopted by 800 million active users, which is roughly 10% of the world’s adult population. The gender gap has narrowed dramatically, and the growth rate is higher in lower-income countries. While there is steady growth in work-related usage, — which  is more common for educated users in higher paid professional occupations, the growth has been even faster in non-work related activities, which grew from 53% to more than 70% of all usage.

    “The data suggest that AI is being democratized faster than previous technologies, with early disparities in geography and gender closing at a remarkable pace,” wrote Friar. Her essay summarized the key changes in the use of ChatGPT over the past three years:

    • Growth in low- and middle-income countries outpaced rich countries by more than four to one over the past year.
    • As of mid-2025, internet-enabled penetration in countries at the 25th percentile of GDP per capita matched adoption in high-income economies.
    • The gender gap closed: users with typically feminine first names slightly outnumbered those with typically masculine names.
    • While all activity grew, non-work usage grew faster than work-related ones, climbing from 45% of traffic in July 2024 to more than 60% by mid-2025.
    • Nearly 80% of conversations involve practical guidance, information seeking, or writing, indicating that people are mostly trying to get something done.
    • Education stands out —10% of all messages, and nearly 40% of practical-guidance queries, are tied to tutoring or teaching.
    • Among work tasks, writing leads the way, with editing and improving existing text outpacing new content creation by two to one, demonstrating that people are using ChatGPT as a tool to augment their work, not do it for them.

    So, will AI democratize intelligence or will it lead to a new kind of AI divide? (more…)

    , , , , , , , ,

  • The 2026 AI Index Report

    1 comment on The 2026 AI Index Report

    A few weeks ago, the Stanford Institute for Human-Centered Artificial Intelligence (HAI) released the 2026 AI Index Report, its ninth 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 offers one of the most comprehensive, data-driven views of the continuing evolution of artificial intelligence.

    “A year ago, this report documented AI’s arrival as a mainstream force,” wrote the report’s co-directors Yolanda Gil and Raymond Perrault. “This year’s data shows what happens after arrival.”

    “This is a technology that has reached mass adoption faster than the personal computer or the internet. Generative AI hit nearly 53% population-level adoption within three years. Leading AI companies are reaching meaningful revenue scale in a fraction of the time it took previous technology generations, and global corporate investment more than doubled in 2025. Organizational adoption rose to 88%, and early estimates suggest the consumer value of generative AI has grown substantially within a year.”

    “The data does not point in a single direction,” they added. “It reveals a field that is scaling faster than the systems around it can adapt.” (more…)

    , , , , , , , , ,

  • Managing AI as Both Tool and Teammate

    No comments on Managing AI as Both Tool and Teammate

    Executives have long relied on simple categories to frame how technology fits into organizations: tools automate tasks, people make decisions, and strategy determines how the two work together. But the rise of agentic AI is beginning to blur these familiar distinctions. As noted in “The Emerging Agentic Enterprise: How Leaders Must Navigate a New Age of AI,” a November 2025 report by the MIT Sloan Management Review in collaboration with the Boston Consulting Group, a new class of systems is emerging that does not fit neatly into traditional management frameworks. The study, conducted by Sam Ransbotham, David Kiron, Shervin Khodabandeh, Sesh Iyer, and Amartya Das, explores how organizations are beginning to grapple with this new reality.

    “That framing is no longer sufficient,” the authors wrote. “A new class of systems — agentic AI — complicates these boundaries. These systems can plan, act, and learn on their own. They are not just tools to be operated or assistants waiting for instructions. Increasingly, they behave like autonomous teammates, capable of executing multistep processes and adapting as they go.” Notably, 76% of respondents to their global executive survey said they view agentic AI as more like a coworker than a tool.

    Agentic AI introduces a fundamentally new category: systems that behave partly like software tools and partly like human coworkers. Traditional management logic assumes that technology either substitutes or complements labor, automates or augments tasks, functions as labor or capital, or behaves as a tool or a worker — but not all at once. Organizations now face an unprecedented challenge: managing a single system that requires both human resource approaches and asset management techniques.

    This dual nature creates a set of tensions that traditional management frameworks were never designed to address. “Technology executives focus on technical issues, making pilot, vendor, or infrastructure decisions. IT leaders want predictable, scalable systems with clear technical specifications. Business executives focus on markets, competition, and people. CFOs need investment models with measurable returns and depreciation schedules. HR executives require performance management frameworks and supervision protocols. Business leaders demand both efficiency and adaptability from the same system.” (more…)

    , , , , , , ,

  • How Is AI Shaping the Future of Work?

    No comments on How Is AI Shaping the Future of Work?

    Earlier this year, the National Academies of Sciences (NAS) released “How Is AI Shaping the Future of Work,” a video podcast and accompanying transcript of a conversation between MIT economist David Autor and writer Sara Freuh in Issues in Science and Technology, a quarterly journal jointly published by the NAS and the University of Arizona.

    “For as long as people have speculated about the development of artificial intelligence, they have debated its potential impacts on the labor market,” noted the article in its introduction. “Today, several years into widespread use of large language models, those questions are more urgent, but the answers are less clear. Is AI already taking jobs away? Could human beings flourish in a world in which they no longer have to perform economically valuable work?”

    “As artificial intelligence becomes ever more advanced and capable, it’s being harnessed more frequently in American workplaces, and it’s raising thorny questions about how it will impact the nature and number of human jobs in the future,” Freuh added. “I’m joined today by David Autor, a professor at the Massachusetts Institute of Technology and one of the world’s leading labor economists. He’s known for his research on the effects of globalization, automation, and technological change on the labor market … and he’s written extensively on the subject.”

    I didn’t focus much on economics during my long career as a technologist and business strategist at IBM. But, since becoming affiliated with MIT in 2005, I’ve become increasingly interested in economics — in particular, on the impact of advanced technologies like the internet and AI on jobs and the labor markets. And, I can personally vouch for Freuh’s comments about Professor Autor’s excellent articles, from which I’ve learned so much over the past couple of decades.

    Let me highlight some of the key themes in the conversation between Freuh and Autor. (more…)

    , , , , , , , ,

  • What Causes AI Brain Fry?

    1 comment on What Causes AI Brain Fry?

    In a recent Harvard Business Review (HBR) article, “When Using AI Leads to ‘Brain Fry,’” participants in a research study conducted by the Boston Consulting Group described experiencing “a mental fog with difficulty focusing, slower decision-making, and headaches” due to excessive use or oversight of AI tools beyond their cognitive capacity. The article’s authors — Julie Bedard, Matthew Kropp, Megan Hsu, Olivia T. Karaman, Jason Hawes, and Gabriella Rosen Kellerman — named this condition “AI brain fry.”

    Their findings are based on a survey of 1,488 full-time U.S.-based workers (48% male, 51% female) at large companies across a variety of industries, roles, and levels. The survey asked about patterns and intensity of AI use, work experiences, and cognitive and emotional effects.

    The authors concluded  that AI brain fry arises not from the technology itself, but from the “excessive use or oversight of AI tools beyond one’s cognitive capacity.”

    “This problem is becoming more common,” the authors wrote. “As enterprises use more multi-agent systems, employees find themselves toggling among more tools. Contrary to the promise of having more time to focus on meaningful work, juggling and multitasking can become defining features of working with AI.” Unsurprisingly, workers are running up against the limits of their cognitive capacity. In recent weeks, online AI users have described increased cognitive load, “saturated” attention, and mental fatigue in social media posts. (more…)

    , , , , ,

  • Know Your Agent (KYA): Governing AI in the Enterprise

    No comments on Know Your Agent (KYA): Governing AI in the Enterprise

    “Agents will be embedded in business processes with direct impact on operations, data, and compliance,” wrote Eric Broda and John Y Miller  in “KYA — Know Your Agent,” a recent article in their Agentic Mesh Substack. Broda is the founder of Broda Group Software, co-founder of The Agentic Mesh Company, and co-author of Agentic Mesh, a book published in February 2026; Miller is an Agentic AI Systems Builder and Podcast Host.

    What Is KYA and How Does It Work?

    In the early days of the digital economy, organizations learned that trust could not be assumed — it had to be engineered. That realization led to the development of frameworks like Know Your Customer (KYC) and Know Your Business (KYB), thus enabling institutions to verify identities, manage risk, and operate at scale in increasingly complex environments. “KYC and KYB exist for a simple reason: when a bank or regulated firm lets someone open an account, move money, or access sensitive services, the firm needs confidence about who is on the other end and what risks they bring.”

    A similar challenge is now emerging, but with a new kind of actor: AI agents. “Agents will be embedded in business processes with direct impact on operations, data, and compliance. This creates three engineering requirements: limit what the agent can do, record what it did, and explain why it did it.”

    In the emerging age of AI, enterprises face a fundamental question: how do you establish trust, accountability, and controls when the “actor” is no longer a person or an organization, but an intangible asset, i.e., software acting with increasing autonomy?

    Know Your Agent (KYA) is one answer to that question. It extends the principles of identity, risk management, and governance from people and institutions to AI agents operating at scale — ensuring that their behavior is transparent, auditable, and aligned with organizational objectives. (more…)

    , , , , , , , , , , , ,