IWB Blog

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

ABOUT

RECENT POSTS

CATEGORIES

  • Should We Fear an AI-Induced Jobs Apocalypse?

    1 comment on Should We Fear an AI-Induced Jobs Apocalypse?

    “The launch of ChatGPT in 2022 ignited the artificial-intelligence boom — and elicited a chorus of warnings from AI bosses of an impending jobs apocalypse,” said The Economist in “Prepare for an AI jobs apocalypse,” the lead article in its May 16 issue. “Never mind that they have reason to talk up the disruptiveness of their products, or that rich-world employment is near all-time highs — the dark message has landed.”

    Public anxiety about AI-driven unemployment has clearly been rising. Seven in ten Americans believe AI will make it harder to find work, while nearly one-third fear for their own jobs. The concerns have been amplified by a slowdown in hiring for recent college graduates, especially in technology-related occupations such as software development.

    Yet, despite these fears, the historical record offers reasons for caution before concluding that AI will trigger mass unemployment.

    “Labour markets constantly change,” noted The Economist. “Today’s offices would be unrecognisable to a worker from 50 years ago. Never in modern history has technological progress hurt the overall demand for human labour.”

    The article points to the so-called Engels pause, the period from roughly 1790 to 1840 during the early Industrial Revolution when steam-powered mechanization transformed manufacturing. During those decades, British GDP expanded rapidly while working-class wages stagnated, fueling social unrest and movements like the Luddites, who destroyed textile machinery they believed threatened their livelihoods.

    But more recent economic history suggests that technology itself was not the primary cause of stagnant living standards during that period. As The Economist noted in a related Money Talks newsletter, historians now attribute much of the hardship to high food prices driven by wars and trade policies rather than to automation directly reducing labor demand.

    This historical perspective matters because it reminds us that technological change has repeatedly disrupted labor markets without causing permanent mass unemployment.

    Americans are quite pesimistic about the impact of AI on jobs

    “At no time in polling history have Americans been less optimistic about their long-term employment prospects,” wrote The Economist in a second article, “The jobs apocalypse: a (very) short history.” According to one survey, the average person believes there is a 22% chance of losing their job within five years — a higher level of anxiety than during the global financial crisis of 2007–09.

    The fears are increasingly tied to AI. A 2025 Gallup survey found that nearly three-quarters of Americans expect A.I. to slash jobs. And, nearly one in five American workers recently told pollsters that AI or automation is “very” or “somewhat” likely to replace them.

    Public anxiety has also been reinforced by statements from prominent AI executives. Anthropic CEO Dario Amodei has warned that AI could potentially drive unemployment rates to 10–20% within five years. Similarly, Sam Altman, CEO of OpenAI, has said that workers should expect significant changes in their roles as AI systems become more capable and widely adopted.

    But economist are far less apocalyptic

    Many economists reject the so-called “lump of labor fallacy” — the mistaken belief that there is a fixed amount of work in the economy. Throughout modern history, technological progress has simultaneously destroyed some jobs while creating new industries, occupations, and forms of work.

    This question has long fascinated economists. If technology has been automating human labor for over two centuries, why hasn’t automation already eliminated most jobs?

    MIT economist David Autor addressed this question in his influential 2015 paper, “Why Are There Still So Many Jobs? The History and Future of Workplace Automation.” The answer, Autor argued, reflects a fundamental economic reality: tasks that cannot be substituted by automation are often complemented by it.

    Automation certainly replaces some human labor. But it also raises productivity, lowers costs, expands demand, and creates opportunities for new kinds of work. Discussions about automation often focus excessively on job substitution while underestimating the complementary relationship between technology and labor.

    Most jobs consist of multiple tasks. Some tasks are relatively easy to automate, while others require judgment, creativity, interpersonal skills, or physical dexterity. Automating routine portions of a job does not necessarily eliminate the entire occupation. In many cases, automation increases the productivity and value of workers by allowing them to focus on the aspects of their jobs where human capabilities matter most.

    Research on technology and employment has often emphasized the jobs lost to automation while paying less attention to the creation of entirely new occupations. To better understand this process, Autor and his collaborators — economists Caroline Chin, Anna Salomons, and Brian Seegmiller — analyzed the emergence of new work in the US economy over the past eight decades in their 2022 paper, “New Frontiers: The Origins and Content of New Work, 1940–2018.”

    Their findings are striking. Roughly 60% of employment in 2018 consisted of occupations introduced since 1940. From 1940 to 1980, much of this new work emerged in mid-skill manufacturing, operations, administrative, and clerical occupations. But from 1980 to 2018, new job creation shifted increasingly toward high-skill managerial, professional, and technical occupations, alongside lower-paid service jobs involving physical and interpersonal tasks such as health aides, food services, and personal care.

    So far, labor-market data does not show evidence of an AI-driven jobs collapse.

    “The labour market certainly is not cracking yet,” said The Economist. Employment rates across OECD countries remain near record highs, unemployment is historically low, and the US continues to add jobs in many AI-exposed industries. The US Bureau of Labor Statistics projects that the economy will add more than five million jobs between 2024 and 2034.

    But history is not always a reliable guide to the future. Today’s AI systems are improving at extraordinary speed. Advanced models can now perform coding, writing, research, and reasoning tasks that seemed out of reach only a few years ago. AI agents are proliferating rapidly, and business investment in AI infrastructure continues to surge.

     There is little direct evidence in labor-market data that AI is destroying jobs at scale. But, given the pace of technological progress, The Economist cautions that it would also be unwise to dismiss the possibility of significant disruptions ahead. The more important question may not be whether AI eliminates work altogether, but how quickly labor markets, institutions, and political systems can adapt to large-scale economic reallocation.

    What should governments do?

    This raises difficult policy questions.

    China, for example, has encouraged companies to deploy AI while discouraging layoffs. Other economists have proposed higher taxes on capital or levies on data centers.

    Some proposals focus on slowing technological adoption. But attempts to significantly inhibit technological progress would likely prove counterproductive. AI has the potential to generate enormous benefits — not only economic growth and productivity gains, but also advances in healthcare, scientific discovery, climate research, and poverty reduction.

    Had the Luddites succeeded in halting industrial automation two centuries ago, the world would almost certainly be far poorer today.

    A more constructive approach would focus on helping workers adapt to economic transitions while ensuring that the gains from AI are more broadly shared. If AI substantially increases profits in sectors such as software, semiconductors, and cloud infrastructure, governments could explore tax reforms that capture a portion of these economic rents more effectively.

    At the same time, stronger labor-market adjustment policies could help workers navigate disruptions. Wage insurance programs could cushion income losses during career transitions, while active labor-market policies — such as Denmark’s highly regarded retraining and job-placement programs — could help workers move into new occupations more quickly.

    These policies would improve economic resilience and fairness regardless of AI’s ultimate impact on employment.

    The jobs apocalypse may never arrive. But if governments wait for definitive evidence before strengthening social protections and workforce transition mechanisms, they may find themselves responding too late to potentially profound economic changes.

    , , , , , ,

  • Higher Ed Is Sending Mixed Signals on AI

    1 comment on Higher Ed Is Sending Mixed Signals on AI

    Earlier this year I wrote a blog about the continuing debate across higher education on whether — and how — university students should be taught to properly use generative AI, “How Should University Students Be Taught the Proper Use of AI.” I mentioned in the blog that the debate about AI reminds me of my own early experiences with computers as a physics student at the University of Chicago in the 1960s — a time when the legitimacy of using computers as a tool for conducting research in physics and other disciplines was also in question.

    At the time, the use of computers for scientific research was still relatively new. My graduate student advisor was professor Clemens Roothaan, one of the pioneers in the use of computers in physics and chemistry research. Some older physics professors looked askance at the growing use of computers, and told me that this wasn’t “real physics” — that is, the kind of pencil-and-paper physics they had grown up with over the past few decades. Their reaction echoes today’s discussions about the proper use of AI-based tools in higher education.

    “As graduates leave our campuses, companies are expecting to welcome a cohort of digital natives who’ll be ready to use AI with confidence,” wrote Miami University management professor Megan Gerhardt in a Harvard Business Impact article, “Higher Ed Is Sending Mixed Signals of AI.” “Instead, they’ll be getting new degree-holders who spent the last four years receiving warnings from at least some of their faculty that AI use is cheating, not a critical skill for their future careers.”

    Gerhardt writes that she sees this gap from two vantage points. As a professor at Miami University where she teaches change management and leadership to undergraduates. And as a consultant and strategist for organizations around the world, for whom she helps reframe their generational challenges as opportunities for innovation and learning.

    “The distance between those two worlds has never felt wider.” As a concrete example of this gap, she cites a faculty meeting earlier this year where several of her colleagues discussed how to shut down the use of generative AI in their classrooms. That same week, one of her corporate clients asked her why their Gen Z interns were reluctant to use their company’s AI platform. “The irony was not lost on me,” she added.

    Students are paying the price for these mixed signals. A recent graduate shared her direct experience: “There was a significant knowledge gap that I only recognized once I began my internships. In many classrooms, AI is stigmatized. However, during my internships, I was strongly encouraged to use it. As a result, I fear that I may struggle to adjust to professional environments.” (more…)

    , , , , , , , ,

  • Proprietary vs Open AI Models: Experiment → Learn → Specialize → Scale

    1 comment on Proprietary vs Open AI Models: Experiment → Learn → Specialize → Scale

    Open AI models present an intriguing economic puzzle. They can often deliver performance approaching that of leading proprietary models at a fraction of the cost. Yet enterprises continue to overwhelmingly favor closed, proprietary models — and by some measures, their share of enterprise AI spending is actually increasing. Why is that the case?

    The answer may have less to do with the relative merits of open versus proprietary AI than with where enterprises currently find themselves on the AI adoption curve.

    “Artificial intelligence is reshaping economic systems at a pace we have rarely seen in modern technological history,” wrote Frank Nagle, — research scientist at MIT’s Initiative on the Digital Economy and newly appointed chief economist at Microsoft’s AI Economy Institute, — in a November 2025 blog, “Revealing the Hidden Economics of Open Models in the AI Era.”

    “Yet amid the excitement, a crucial part of the story has been missing,” he added. “Specifically, understanding the role that open models play in the AI economy, and how much value is being left on the table when organizations overlook open alternatives, are two topics requiring a closer look.”

    In “The Latent Role of Open Models in the AI Economy,” Nagle and Georgia Tech professor Daniel Yue probed these questions by analyzing a comprehensive dataset of AI model usage, prices, and performance. Their research uncovered a striking economic puzzle.

    Closed models account for roughly 80% of model usage and 96% of revenue, even though they cost, on average, six times more than competing open models. Meanwhile, open models routinely achieve 90% or more of the performance of proprietary models and typically close much of the performance gap within a few months after the release of a new closed model.

    “The findings surprised even us,” Nagle wrote.”

    So why are enterprises continuing to spend so much more on proprietary models?

    A recent article, “Everyone Is Wrong About Open Source AI in the Enterprise,” by Decagon co-founder and CEO Jesse Zhang, offers a very interesting explanation.

    “The prevailing story right now is that open source is eating the enterprise,” Zhang wrote. “The capability gap between the best closed and open models has shrunk to low single digits. A third of the Fortune 500 has verified accounts on Hugging Face.”

    “And yet enterprise spend as a whole is moving the opposite direction. Open source models just fell to 11% of enterprise LLM spend, down from 19% a year ago. The trend is actually moving the other way compared to the popular narrative. Why is this and what does it mean for the future?”

    Frontier Models for Discovery; Open Models for Production

    Zhang offers a compelling resolution of this apparent paradox based on his own company’s experience. Decagon develops conversational AI agents for customer service, and roughly 90% of its workloads run on open models. “It wasn’t cost, and it wasn’t because our customers demanded it (though they don’t mind it),” he wrote. “It was because we had no other option.”

    The reason has to do with where an AI application is in its development lifecycle.

    “When a use case is new, you want the smartest general-purpose model you can get,” Zhang explained. “You don’t know the shape of the problem yet, so you pay a premium for intelligence you may not end up needing. That’s the right trade at that stage.”

    But the economics change as the application matures.

    “Once the use case is fully built out, when you know the distribution of inputs, the behaviors you need, and the failure modes to guard against, the trade flips. Now general intelligence is overhead, and you want the smallest, fastest model fine-tuned to do your specific thing extremely well.”

    In other words, the AI deployment lifecycle might be summarized as:

    Experiment → Learn → Specialize → Scale

    During experimentation, the most capable frontier models offer significant advantages. Enterprises are still discovering what AI can do, which applications create real value, and how their workflows need to change. At this stage, paying a premium for the flexibility and broad intelligence of a proprietary frontier model can make considerable sense.

    But once an application is well understood and operating at scale, the requirements change. Speed, latency, customization, control, and cost become increasingly important. At that point, smaller, specialized open models may become much more attractive.

    Customer service is a particularly good example.

    “Customer service happens to be one of the most obvious AI use cases in the industry,” Zhang wrote. It involves “well-understood workflows, enormous conversation volume, tight quality bars.”

    Companies like Decagon may simply be further along the AI maturity curve than the average enterprise deployment. “When you’re running AI agents in production for customer service, latency makes or breaks the product,” Zhang explained. “A conversation where every turn takes 8 seconds is not a product anyone will use.”

    That creates a need for smaller, faster models. But small models generally don’t meet demanding enterprise quality standards without significant customization. “They only get there through heavy fine-tuning on the exact task.”

    And that, according to Zhang, is where open models have an important advantage.

    “The frontier labs don’t really sell this combination. You can’t fine-tune their best models the way we need to, and their small models aren’t ours to shape. Small + fine-tuned means open weights.”

    “The cost savings are real but secondary,” he added, “and enterprise comfort with self-hosted models is a nice side effect, not the reason.”

    This suggests that the continued dominance of proprietary models doesn’t necessarily mean that open models are losing the competition. It may simply mean, as Zhang argues, that “enterprise AI as a whole is at the very beginning of the maturity curve.”

    Enterprise AI Moves at Enterprise Speed

    This argument is consistent with a theme I’ve discussed in several previous blogs: AI technology may be advancing at extraordinary speed, but enterprise adoption is moving much more slowly.

    In a September 2025 blog, “The AI Revolution Will Happen in Enterprise Time,” I referenced a Goldman Sachs article, “The Outlook for AI Adoption as Advancements in the Technology Accelerate,” which noted that despite high expectations among executives, “AI is changing business much more slowly than expected.” While technical achievements are racing ahead, enterprise adoption still has to overcome corporate inertia and demonstrate a convincing return on the enormous investments being made.

    An article in The Economist similarly asked, “Why is AI so slow to spread?” “Talk to executives and before long they will rhapsodise about all the wonderful ways in which their business is using artificial intelligence.” But announcing AI projects and fundamentally transforming business processes are two very different things.

    I also discussed this issue in a subsequent blog about “AI as a Normal Technology,” a research paper by Princeton computer science professor Arvind Narayanan and PhD candidate Sayash Kapoor.

    They argue that, as with earlier general-purpose technologies, the economic and societal impact of AI will unfold gradually because “diffusion is limited by the speed of human, organizational, and institutional change.”

    The ultimate impact of AI won’t be determined simply by how rapidly model capabilities improve. It will depend on how quickly those capabilities are translated into useful applications and diffused throughout organizations and the broader economy.

    This helps explain why proprietary models may dominate enterprise AI today even if open models eventually become much more important.

    Most enterprises are still experimenting. They are trying different models, identifying promising use cases, redesigning workflows, and figuring out where AI can deliver measurable value. They often haven’t yet accumulated the experience, data, infrastructure, or specialized expertise required to optimize and fine-tune smaller open models.

    At this stage of the maturity curve, it’s simply easier to plug into a highly capable proprietary frontier model. Organizations don’t have to manage the underlying infrastructure, and they retain the flexibility to experiment and iterate as their understanding of the application evolves.

    What Happens as Enterprise AI Matures?

    Zhang believes the balance will eventually shift.

    “If that’s right, then every use case being prototyped on a frontier model today is a future open source migration,” he wrote. “As deployments mature, companies will do what we did: distill, fine-tune, specialize. The frontier labs will keep owning discovery. Open source will increasingly own production.”

    It’s a compelling hypothesis, but the outcome isn’t necessarily inevitable.

    Proprietary AI providers aren’t standing still. They can lower prices, offer smaller and faster models, expand fine-tuning and customization capabilities, and provide managed infrastructure that many enterprises may prefer to operating their own open AI systems.

    And, as Zhang himself acknowledges, fine-tuning takes considerable effort, and many organizations don’t yet have the necessary resources or expertise. A use case generally needs to be sufficiently valuable, stable, and deployed at enough scale for the investment to make sense. Organizations also need enough high-quality data to ensure that a smaller specialized model can perform as well as a frontier model on the particular task.

    Until those conditions are met, using a proprietary frontier model may remain the simpler and more economically sensible choice. Zhang therefore expects open models’ share of enterprise LLM spending eventually to rise — but not necessarily anytime soon.

    The future may not belong exclusively to one or the other.

    Enterprises may increasingly use different models for different stages and purposes — powerful frontier models to explore new applications and specialized open models where workloads are stable, scale is large, and customization, latency, control, or economics justify the additional investment.

    The result could be a heterogeneous enterprise AI environment in which proprietary and open models coexist, with organizations selecting the right model for the particular problem they are trying to solve, as is the case with open source software (OSS), where despite major advances over the past three decades, OSS coexists with proprietary software.

    Today’s overwhelming share of enterprise spending going to proprietary models may tell us less about the ultimate competitive position of open AI than about the still-early maturity of enterprise AI itself.

    The history of enterprise technology suggests that the transition from experimentation to large-scale production takes time. As AI follows that same path, the balance between proprietary and open models will likely change with it.

    , , ,

  • AI Moves Toward Organizational Deployment

    No comments on AI Moves Toward Organizational Deployment

    “Much of the effort and attention around AI for the last several years has been around technical developments,” wrote Babson College professor Tom Davenport in a recent Substack post, “The AI World Moves Toward Organizational Deployment.” “New model announced! New benchmark surpassed! New contract for massive data centers! New world-class technologists hired! You know the drill.”

    “I am happy to say, however, that things are beginning to change,” he added. “AI companies are beginning to realize something that many corporate executives knew intuitively. What matters isn’t the technology — OK, that’s important too — but the ability of organizations to deploy it effectively and get value from it.”

    That observation captures what I believe is one of the most significant developments in AI today. For the past several years, the conversation has focused overwhelmingly on increasingly powerful models and ever larger computing infrastructures. Those advances remain essential, but they are no longer sufficient. As AI matures, the central challenge is becoming organizational rather than technological: How do companies redesign work, management, and decision making to capture AI’s potential? (more…)

    , , , , , ,

  • Why AI Hasn’t Replaced Software Engineering? — and Probably Won’t

    No comments on Why AI Hasn’t Replaced Software Engineering? — and Probably Won’t

    A couple of weeks ago I posted “Why AI Can’t Replace Software Engineering,” a blog in which I discussed an important distinction that is often overlooked: software engineering is not synonymous with programming.

    I keep coming back to this topic because, in my opinion, the difference between programming and software engineering offers a particularly concrete way of understanding the difference between human and AI capabilities. It is also highly relevant to current debates about jobs and automation, since several widely discussed studies have identified programming as one of the occupations most exposed to AI.

    Given the extraordinary advances in AI coding tools, why haven’t software engineering jobs already begun to disappear?

    I recently read an excellent essay, “Why AI hasn’t replaced software engineers and won’t,” by Princeton computer science professor Arvind Narayanan and PhD candidate Sayash Kapoor that sheds considerable light on this question.

    “There is great anxiety and uncertainty about AI replacing jobs,” wrote the authors. “How can we move past vague warnings and bombastic predictions and bring data to bear on this question? One good way is to look at the profession where AI capabilities are furthest along and adoption has been exceptionally rapid: software engineering.”

    They argue that there is already enough evidence to reject the narrative that once AI reaches some capability threshold, mass layoffs will inevitably follow. If software engineering — a profession with relatively few regulatory barriers and where AI adoption has been extraordinarily rapid — has not experienced widespread labor displacement, then many other professions are likely to be even more insulated. (more…)

    , , , , , , , ,

  • Some Personal Reflections on a Long Career

    2 comments on Some Personal Reflections on a Long Career

    On May 5, 2012, I gave the commencement address at Penn State’s College of Information Sciences and Technology. Then Dean David Hall offered me excellent advice on what makes for a good commencement speech: make it personal, tell us about yourself and your background, share a few lessons you’ve learned over the years, and keep it short.

    I followed his advice and reflected on my 60-year involvement with computers and the IT industry. Looking back today, more than a decade later, those reflections still ring true. If anything, the recent AI-driven technological changes have reinforced the lessons my career has taught me about curiosity, adaptability, and the importance of embracing unexpected opportunities.

    I started my talk by discussing the serendipitous nature of how I first got involved with computers. (more…)

    , , , ,

  • Why AI Can’t Replace Software Engineering

    No comments on Why AI Can’t Replace Software Engineering

    AI’s potential threat to programming has been widely debated over the past few years, as evidenced by the growing number of provocative headlines. For example, “The End of Programming” argued that “the end of classical computer science is coming, and most of us are dinosaurs waiting for the meteor to hit.” Other articles have predicted that “ChatGPT Will Replace Programmers Within 10 Years,” while NVIDIA CEO Jensen Huang has suggested that coding may no longer be a viable long-term career because of increasingly capable AI systems.

    In addition, several widely discussed studies have identified programming as one of the occupations most exposed to AI and have documented declining employment among younger workers in software-related jobs. Unsurprisingly, parents of computer science students increasingly worry about their children’s career prospects.

    But these concerns raise a more fundamental question: What exactly is being threatened? Is AI making software engineering obsolete, or is it primarily changing the nature of programming?

    “Social media provide a steady diet of dire warnings that artificial intelligence (AI) will make software engineering (SE) irrelevant or obsolete,” wrote Carnegie Mellon computer scientists Mary Shaw and Eunsuk Kang in their 2024 article “Chill, Y’all: AI Will Not Devour SE.” “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.”

    I have long known professor Shaw since we both served in the President’s Information Technology Advisory Committee (PITAC) in the late 1990s. She’s been writing about software as an engineering discipline, going back to her 1990 article on “Prospects for an Engineering Disciplines of Software.” Over the years, I have closely followed her continuing research on the topic.

    “Software engineering (SE) is a rich, robust discipline that covers software systems from idea through their lifetime,” Shaw and Kang explained in their article. SE is “the branch of computer science that creates practical, cost-effective solutions to computing and information processing problems, by applying the best-systematized knowledge available, developing software systems in the service of mankind.” It “encompasses the full scope of software systems from concept to retirement — a full spectrum of issues from understanding what problem the software should solve through overall design, tradeoff resolution, performance, reliability, sustainability, usability, fitness for purpose, programming of components, composition of components, validation, adherence to policy and standards, and evolution.” (more…)

    , , , , , , ,

  • The 2026 State of Tech Talent

    No comments on The 2026 State of Tech Talent

    The “2026 State of Tech Talent Report” was recently published by Linux Foundation Research. “Over the past several years, the Linux Foundation has surveyed hiring and training stakeholders to capture the state of the technical talent market amid technological shifts and economic changes,” wrote authors Adrienn Lawson, Marco Gerosa, and Anna Hermansen in the report’s executive summary. “This year’s study is based on an online survey fielded in February 2026, which collected responses from 400 participants worldwide and examined the impact of AI, especially generative AI, on the IT talent market.”

    The report’s central conclusion is both clear and timely: despite widespread concerns about AI-driven job displacement, the technology sector is facing a skills crisis rather than a jobs crisis.

    “While AI is a net driver of job creation in IT, with a +31% net hiring effect expected for 2026, organizations are struggling with a major full-stack readiness problem,” wrote the authors. “Security concerns are also the #1 barrier to getting value from new technologies. To counter these challenges, organizations strongly prefer upskilling existing staff over external hiring, preserving institutional knowledge and boosting retention.”

    “There is little doubt that AI is changing the IT job market,” the report added. “Adoption is widespread and organizations across industries and regions are integrating AI into core business functions.” This widespread adoption raises important questions. Where do organizations expect AI to deliver value? Is that value coming at the expense of technical jobs, or is it creating new ones? What skills are required to operationalize AI at scale, and how are organizations developing those capabilities? (more…)

    , , , , , , ,

  • The State of AI in Higher Education: Widespread Use, Unclear Rules

    No comments on The State of AI in Higher Education: Widespread Use, Unclear Rules

    “Artificial intelligence is shaping how many college students think about their academic paths,” said a recently published Gallup article, “College Students Weigh AI’s Impact on Majors and Careers.” Forty-two percent of bachelor’s degree students say AI has caused them to give serious thought to changing their major, including 13% who say they have thought about it a great deal. Even more associate degree students — 56% — say AI has prompted them to reconsider their field of study.

    These findings reflect a broader reality: AI is already deeply embedded in higher education, even as many colleges and universities are still struggling to determine how students should use it and how faculty should teach it.

    The Gallup article is based on a web survey conducted in October of 2025 by Gallup and the Lumina Foundation among nearly 4,000 US college students pursuing bachelor’s and associate degrees. The results were published in a joint Lumina-Gallup report on AI in higher education.

    The survey found that 57% of students already use AI daily or weekly for schoolwork, while only 13% say they never use it. Yet more than half of students report that their institution discourages or outright prohibits AI use in coursework, and 52% say that at least some of their classes lack clear guidance on acceptable AI use.

    The findings point to a widening gap between student behavior and institutional policy — one with significant implications for academic integrity, teaching practices, and workforce preparation. (more…)

    , , , , , , , ,

  • 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…)

    , , , , , , ,