Follow the emerging research and new practices of using AI at work

With Ajay Agrawal and Joshua Gans he wrote 'Prediction Machines' and 'Power and Prediction', which describe AI as a drop in the cost of prediction; his 2026 paper 'O-Ring Automation' (with Gans) argues AI's job effects depend on bottleneck tasks, not average task exposure.

Co-authored (with Gans and Goldfarb) 'Artificial Intelligence: The Ambiguous Labor Market Impact of Automating Prediction' (Journal of Economic Perspectives, 2019), which treats AI as a drop in the cost of prediction and argues its effect on jobs can go either way; also co-author of the books Prediction Machines and Power and Prediction.

Co-authored (with Agrawal and Goldfarb) 'The Turing Transformation: Artificial Intelligence, Intelligence Augmentation, and Skill Premiums' (Harvard Data Science Review), which looks at how AI that augments workers can change the pay gap between higher- and lower-skilled workers; writes regularly about AI economics on his Substack.

In 'Generative AI at Work' (QJE 2025, with Danielle Li and Lindsey Raymond) he found an AI assistant raised customer-support agents' productivity by about 14 percent on average, with the largest gains for newer, less-skilled workers; his 'Canaries in the Coal Mine' work tracks falling employment for young workers in AI-exposed jobs.

Co-author of 'How People Use ChatGPT' (NBER WP 34255, Sept 2025), which found non-work messages grew from 53% to over 70% of ChatGPT use, with practical guidance, information seeking and writing making up nearly 80% of conversations.

Co-authored 'Algorithmic Bias? An Empirical Study of Apparent Gender-Based Discrimination in the Display of STEM Career Ads' (won the 2026 Donald G. Morrison Long-Term Impact Award), a study of why an ad-delivery algorithm showed STEM job ads unevenly to men and women.

'Robots and Jobs: Evidence from US Labor Markets' (JPE 2020, with Pascual Restrepo) found each additional industrial robot per thousand workers lowered local employment and wages, and 'The Simple Macroeconomics of AI' (2024) estimated modest AI productivity gains over a decade.

'New Frontiers: The Origins and Content of New Work, 1940-2018' (QJE 2024) showed that most of today's jobs are in occupations that did not exist in 1940, and his 2026 patent-drafting field experiment tests whether AI help builds or erodes worker expertise.

Co-author with Daron Acemoglu of 'Power and Progress: Our 1,000-Year Struggle Over Technology and Prosperity' (2023), a history of how major technology shifts have played out for workers and the wider economy, used as a frame for thinking about AI.

Co-author of 'GPTs are GPTs' (Science 2024), which estimated that about 80 percent of US workers have at least 10 percent of their tasks exposed to large language models, and of 'The Productivity J-Curve' (AEJ: Macro 2021) on why new technologies at first depress measured productivity.

'Scenarios for the Transition to AGI' (2024, with Donghyun Suh) models what happens to output and wages if AI can eventually do all human work, and his 'Generative AI for Economic Research' (JEL 2023) showed economists how to use language models in research.

'The Rapid Adoption of Generative AI' (with Bick and Blandin) used a national survey to show US workers took up generative AI faster than the PC or internet, and he co-authored OpenAI's 'How People Use ChatGPT' (2025).

With Acemoglu he wrote 'Robots and Jobs' (JPE 2020) and 'Tasks, Automation, and the Rise in U.S. Wage Inequality' (Econometrica 2022), which attributes a large share of the rise in US wage inequality since 1980 to automation displacing workers from their tasks.

Lead author of 'GPTs are GPTs' (arXiv 2023; Science 2024), the widely used measure of which occupations' tasks large language models could speed up, later extended to firms (AEA P&P 2025).

Co-author of 'Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence' (2025, updated August 2026), which found employment for 22- to 25-year-olds in the most AI-exposed jobs is about 19 percent below comparable peers, driven by less hiring rather than more layoffs.

Co-author of 'Generative AI at Work' (QJE 2025), the large field study showing an AI assistant raised call-center agent productivity most for novice workers, and of 'Hiring as Exploration', which showed a hiring algorithm designed to explore picked more diverse and higher-quality candidates.

First-hand data work behind 'Generative AI at Work' (QJE 2025, with Brynjolfsson and Li), which found the biggest AI productivity gains went to the least experienced customer-support agents; she now studies how algorithms reshape market competition.

Co-authored 'Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence' with Brynjolfsson and Chandar, which uses large-scale payroll data and finds no sign of economy-wide job loss but a clear employment drop for workers aged 22-25 in AI-exposed jobs compared with less-exposed peers.

Co-author of 'AI and jobs. A review of theory, estimates, and evidence' (arXiv 2025, with ILO economists), which finds productivity gains from generative AI are sizable but depend on context, and that demand for novice jobs has fallen in AI-related work; also studied how ChatGPT's release cut demand for writing and translation freelancers.

Lead author of the ILO research brief reviewing evidence on GenAI and jobs (June 2026), which finds productivity gains are real but uneven and often unverified, large-scale job loss has so far been limited, and flags widening gaps, weaker prospects for young workers and changes in how work is organised.

Co-authored 'The Fall of the Labor Share and the Rise of Superstar Firms' (QJE 2020, with Autor, Dorn, Katz and Patterson), which studies how the growth of dominant, highly productive firms relates to the falling share of income paid to workers.

Co-author of the ILO's 2023 study of generative AI and jobs, which concludes the main effect is more likely to augment than automate jobs, with greater exposure in higher-income countries and for women because of their concentration in clerical work.

Lead author of the ILO's global index of occupational exposure to generative AI (2023, updated 2025), which finds about one in four workers worldwide are in jobs with some GenAI exposure and that transformation of jobs is more likely than outright replacement.

Co-authored 'Experimentation and Startup Performance: Evidence from A/B Testing' (NBER WP 26278, 2019, with Hasan and Chatterji), which studies how startups that adopt A/B testing tools perform; his generative-AI papers could not be verified this session.

'Large Language Models, Small Labor Market Effects' (2025, with Emilie Vestergaard; now retitled 'Still Waters, Rapid Currents'), using Danish records linked to worker surveys, found that chatbot adoption had close to zero effect on earnings or hours in the first years.

Known for research on remote and 'work-from-anywhere' arrangements; specific papers and co-authors could not be verified this session, so no finding is stated.

Co-author of the 'Jagged Frontier' and 'Cybernetic Teammate' experiments and of the 'Cyborgs, Centaurs and Self-Automators' study, which sorts professionals into three distinct ways of working with generative AI.

Co-authored 'Algorithms at Work: The New Contested Terrain of Control' (2020) and the 'Jagged Frontier' and 'Novice Risk Work' studies; her 2026 paper with Wiesenfeld and Karunakaran finds organizations scale generative AI when they build support structures for employee experimentation.

Senior author on the 'Jagged Frontier', 'Cybernetic Teammate' and 'Crowdless Future?' (Organization Science 2024) studies, and co-author with Marco Iansiti of 'Competing in the Age of AI' (2020).

Co-author of the 'Jagged Technological Frontier' field experiment with 758 BCG consultants, which found GPT-4 raised speed and quality on tasks inside the AI's capability range but lowered correctness on a task outside it; also writes the widely read One Useful Thing newsletter and the book Co-Intelligence (2024).

Lead author of both the 'Jagged Technological Frontier' (Organization Science 2026) and 'Cybernetic Teammate' (Organization Science 2026) field experiments; the latter found that individuals using AI at Procter & Gamble matched the output quality of two-person teams without AI.

Co-author of 'The Effects of Generative AI on High-Skilled Work' (Management Science 2026), three field experiments at Microsoft, Accenture and a Fortune 100 firm that found access to GitHub Copilot raised completed tasks by about a quarter, mostly for less experienced developers.

Lead author of 'The Impact of AI on Developer Productivity: Evidence from GitHub Copilot' (2023), a controlled experiment in which developers with Copilot finished a coding task 55.8% faster.

Second author of the 'Jagged Frontier' experiment and co-author of 'The GenAI Wall Effect' (2025), which studies the limits of using generative AI to let people outside an occupation do insiders' work.

Co-author of the 'Cybernetic Teammate' experiment and a leading researcher on management practices and reskilling (HBR 'Reskilling in the Age of AI', 2023); her newer work studies how tasks inside firms change after technology adoption.

Lead author of 'The Narrative AI Advantage?' field experiment, which found evaluators given AI-written reasoning rejected startup proposals 9% more often and tended to defer to the AI even when its reasoning was flawed; co-author of 'The Crowdless Future?' (2024).

Co-authored with Whitney Zhang 'Experimental Evidence on the Productivity Effects of Generative Artificial Intelligence' (Science, 2023), which found ChatGPT cut time on professional writing tasks and raised quality, with the biggest gains for weaker writers.

Co-authored with Shakked Noy the 2023 Science paper showing ChatGPT raised productivity on writing tasks and narrowed the gap between stronger and weaker workers.

He led the study that sorted consultants' generative-AI use into 'cyborgs', 'centaurs' and 'self-automators' and traced what each pattern means for building skills.

His book Inside the Invisible Cage: How Algorithms Control Workers (2024; 2025 George R. Terry Book Award) and his 2021 ASQ paper show how an online labor platform's opaque rating algorithm kept freelancers guessing and shaped their behavior.

Wrote Superminds (2018) on groups of people and computers thinking together, and co-authored the 2024 Nature Human Behaviour meta-analysis finding that human-AI combinations on average did worse than the better of humans or AI alone, with gains mostly on content-creation tasks.

Co-authored 'Algorithms at Work: The New Contested Terrain of Control' (Academy of Management Annals, 2020, with Katherine Kellogg and Angèle Christin), the widely cited review of how employers use algorithms to direct, evaluate and discipline workers, and earlier built 'flash organizations' of crowd workers.

As a BCG leader he co-authored the 'Jagged Technological Frontier' field experiment with BCG consultants, which showed AI helps on some tasks and hurts on others that look similar.

His ASQ paper 'Shadow Learning' (2019) showed that surgical trainees shut out by robotic surgery built skill through unapproved practices, and his book The Skill Code (2024) argues intelligent machines are cutting the expert-novice bond that builds human skill.

His 2018 essay with Pachidi and Sayegh on learning algorithms set an early agenda for how machine learning changes expertise, jobs and coordination.

Studies ride-hail and other gig workers under algorithmic management; 'The Making of the Good Bad Job' (2024) argues that apps win workers' consent by giving them constant but tightly limited choices.

Studies how digital tools and AI change the way organizations work; co-wrote The Digital Mindset (2022, with Tsedal Neeley) and 'Pace Layering as a Metaphor for Organizing in the Age of Intelligent Technologies' (Journal of Management Studies, 2022, with Matt Beane).

Her field study of radiologists showed that AI tools are often trained and judged against 'ground truth' labels that leave out how experts actually reason, and a follow-up showed how doctors deal with opaque AI in diagnosis.

Lead author of 'Artificial Intelligence at Work: An Integrative Perspective on the Impact of AI on Workplace Inequality' (Academy of Management Annals, 2025), and studies how platforms and social media change accountability for professionals, e.g. 'Crowd-Based Accountability' (Organization Science, 2022).

Lead author of 'When combinations of humans and AI are useful: A systematic review and meta-analysis' (Nature Human Behaviour, 2024), which found human-AI teams on average did worse than the best of humans or AI alone, doing better on creation tasks than on decision tasks.

Builds tools (Empirica) and 'integrative experiment' designs for large online studies of group decisions and teamwork, and co-authored the 2024 Nature Human Behaviour meta-analysis on when human-AI combinations help and 'The Task Space' framework for team research (2025).

He co-led the Procter & Gamble field experiment ('The Cybernetic Teammate') testing whether individuals using generative AI can match the output of human teams.

Co-discovered a measurable 'collective intelligence' factor in groups, and now studies how AI changes it, e.g. 'Articulating the Role of Artificial Intelligence in Collective Intelligence: A Transactive Systems Framework' (with Pranav Gupta, 2021) and the COHUMAIN human-machine teaming program (2023).

Her ASQ study of investment bankers showed how 'black boxing' algorithmic analysis tools changes who understands and controls knowledge work, and with Matt Beane she described how senior staff learn new technologies from juniors ('inverted apprenticeship').

Her study of a telecom firm's shift to algorithmic sales analytics ('Make Way for the Algorithms') showed how an algorithm-driven way of knowing displaced salespeople's own expertise.

Her ethnographic work shows how people push back on and reinterpret algorithms at work, and her review with Kellogg and Valentine, 'Algorithms at Work', framed algorithmic control as a new contested terrain.

She developed the sociomaterial view of technology at work and, with Susan Scott, showed how TripAdvisor's ranking algorithm and online reviews reshaped the hotel industry.

Co-authored 'Artificial Intelligence and Management: The Automation-Augmentation Paradox' (Academy of Management Review, 2021, with Sebastian Krakowski), which argues that automating and augmenting managerial work cannot be treated as separate choices, and more recently 'Managing with Artificial Intelligence: An Integrative Framework' (Academy of Management Annals, 2025).

An ethnographer of occupations, she co-wrote 'Collaborating with AI' on taking a whole-system view of how AI changes work, and with Gerald Davis examined resistance to algorithmic management of science after generative AI.

He uses machine learning on workplace language (emails, messages) to measure cultural fit and team dynamics, and with Phanish Puranam argues managers' job shifts from prediction to purpose as AI takes over prediction.

Treats human-AI decision-making as an organization design problem ('Human-AI collaborative decision-making as an organization design problem', 2020) and wrote Re-Humanize (2025) on building human-centric organizations as algorithms spread.

He measures collective intelligence in groups and studies how AI teammates change the way people in human-AI teams think and coordinate.

His work with Huggy Rao on organizational friction ('The Friction Project') is about removing bad obstacles in how work gets done, and he now writes about AI in management and teaching (HBR 'The 5 AI Tensions Leaders Need to Navigate'; Substack).

Economist behind much of Microsoft's measurement of AI at work: co-author of 'Working with AI' (2025), 'Shifting Work Patterns with Generative AI' (2025), which found Copilot users spent about two fewer hours a week on email, and lead editor of the New Future of Work Report 2025.

Lead author of the CHI 2024 best paper arguing that generative AI places heavy metacognitive demands on users, and co-author of the CHI 2025 critical-thinking survey of knowledge workers.

Co-leads Microsoft's Tools for Thought work on keeping people's own thinking engaged when they use generative AI, including the CHI 2025 survey of knowledge workers on critical thinking.

Co-authored work showing that generative AI puts heavy demands on users' ability to monitor and steer their own thinking, and on the 'ironies' by which AI assistance can cut productivity.

Senior author of Microsoft's 'Working with AI: Measuring the Applicability of Generative AI to Occupations' (2025), which scored occupations by how much of their work matches what people actually do with Bing Copilot, and co-author with Mary L. Gray of the book Ghost Work (2019) on hidden on-demand labor behind AI.

Runs randomized experiments on how people use AI tools, for example showing how LLM-based search changes decision speed, accuracy and overreliance (CHI 2025).

Co-author of Microsoft's CHI 2025 survey of knowledge workers, 'The Impact of Generative AI on Critical Thinking', which found that higher confidence in GenAI goes with less critical thinking while higher self-confidence goes with more.

He studies how knowledge workers use generative AI in data analysis and spreadsheets, including a participatory prompting study of AI-assisted data analysis and prompt-refinement controls for comprehension tasks.

Co-led Microsoft's early synthesis of Copilot productivity studies (2023) and the 2024 'Generative AI in Real-World Workplaces' report.

Co-authored the NeurIPS 2025 paper that used real Bing Copilot conversations to measure which work activities and occupations AI applies to.

Studies how people relate to each other through work technology; co-edited Microsoft's New Future of Work reports and co-wrote research on video-meeting fatigue during the pandemic.

Lead author of Microsoft's 'Working with AI: Measuring the Applicability of Generative AI to Occupations' (2025), which used real Bing Copilot conversations to score how much each occupation's work activities overlap with what people successfully do with AI.

She built CoAuthor, a dataset of people writing with GPT-3 for studying human-AI co-writing, and her group now runs controlled experiments on how AI use affects critical thinking, reading and writing.

Co-led the large field experiment across many firms showing how access to Microsoft's AI tool changed knowledge workers' time on email and other work (NBER 2025).

Leads the Anthropic Economic Index work and co-wrote 'Labor Market Impacts of AI: A New Measure and Early Evidence' (March 2026), which introduced an 'observed exposure' measure and found no broad rise in unemployment among highly exposed workers since late 2022, though hiring of younger workers in exposed jobs has slowed.

Leads Microsoft's New Future of Work initiative, whose yearly New Future of Work Report (2023, 2024, 2025) summarizes Microsoft's research on how generative AI is changing work, and co-authored the 2024 synthesis 'Generative AI in Real-World Workplaces'.

Economist who co-edited the 2024 and 2025 New Future of Work reports and co-authored a 2026 field experiment on structured protocols for working with AI.

Leads Anthropic's economic research and is first author of recent Anthropic Economic Index reports on how people use Claude for work, plus a 2026 paper proposing a new measure of AI's labor-market impact.

Designs and tests just-in-time wellbeing and stress-reduction interventions for workers, such as her CHI 2022 study on workplace stress-reduction systems.

Lead author of Clio (2024), Anthropic's privacy-preserving system for analyzing real Claude conversations, and co-author of 'Which Economic Tasks are Performed with AI?' (2025), the first Anthropic Economic Index paper mapping Claude usage onto occupational tasks.

Studies how AI changes knowledge work inside organizations, including research on risks to workers from AI-mediated enterprise knowledge access (FAccT 2024).

Her lab builds interfaces that help people notice and check what an AI model chose, such as ChainForge, an open-source visual toolkit for comparing and testing LLM prompts and outputs.

Anthropologist and MacArthur Fellow (2020) who co-wrote Ghost Work (2019) with Siddharth Suri on the hidden on-demand workers behind AI systems, and now studies the human labor of AI red-teaming and content work.

Senior author of 'Generative Agents: Interactive Simulacra of Human Behavior' (2023) and its 2024 follow-up on LLM agents built from interviews that simulate real individuals, which opened the line of work on AI agents that stand in for people; he also studies when explanations reduce over-reliance on AI.

Lead author of OpenAI's 'How People Use ChatGPT' (NBER working paper 34255, September 2025), a large study of how consumers use ChatGPT, including how much of that use is work-related.

With Harvard's Laboratory for Innovation Science (Karim Lakhani, Jacqueline Lane) he ran experiments comparing generative AI with crowds on creative problem-solving and on evaluating early-stage innovations.

He is lead author of the CHI 2025 survey of knowledge workers with Microsoft Research showing that higher confidence in generative AI went with less critical thinking, while higher self-confidence went with more.

Senior author of Stanford's 'Future of Work with AI Agents' (2025), which built the WORKBank database of worker preferences and expert ratings across occupational tasks to show where workers want AI to automate versus augment their work.

His CHI 2024 paper names the 'gulf of envisioning', the gap between what a person wants and how they must phrase it for an LLM, and his CHI 2025 study looks at how software teams prototype with prompts.

Lead author of 'To Trust or to Think' (CSCW 2021), which showed that 'cognitive forcing' interface designs reduce people's over-reliance on AI suggestions, and now studies worker-centric AI for decision support.

Her long-running field studies of attention switching and interruptions in screen-based work, summarized in her book Attention Span, give a baseline for judging how AI tools add to or reduce the attention cost of knowledge work.

Co-author with Ethan Mollick of widely used guides on using AI in teaching ('Assigning AI', 2023; 'Instructors as Innovators', 2024) and of the Generative AI Labs' 'Prompting Science Reports' and the 'Cybernetic Teammate' field experiment on AI and teamwork.

Co-author with Sayash Kapoor of the book 'AI Snake Oil' (2024) and the essay and newsletter 'AI as Normal Technology', which argue AI will spread through the economy slowly like past general-purpose technologies rather than as sudden superintelligence.

Co-author of 'AI Snake Oil' and 'AI as Normal Technology' with Arvind Narayanan, and leads agent-evaluation work (HAL: Holistic Agent Leaderboard, CORE-Bench, CRUX open-world evaluations) that tests how reliable AI agents are at real tasks.

Co-founder of DORA and co-author of 'Accelerate' and the SPACE and DevEx frameworks for measuring developer productivity; her book 'Frictionless' (with Abi Noda, 2025) applies this to removing friction in the AI era.

His book 'The Exponential Age' (2021) and the weekly Exponential View newsletter give business leaders data-driven readings of AI adoption and its economic effects, including tracking of how companies report AI use.

A physician-leader who coined the term 'hospitalist', he explains to clinicians and the public how AI is changing medical work, most recently in his 2026 book A Giant Leap.

His book 'Blood in the Machine' (2023) uses the Luddites to frame today's worker resistance to automation, and his newsletter series 'AI Killed My Job' collects workers' own accounts of losing work to AI.

Co-author of 'Observability Engineering' (2nd ed., 2026), which argues teams shipping faster with AI need faster production feedback loops, and writes on norms for AI use at work, e.g. 'Confessions of an Unrepentant Slop Snob' (2026).

His newsletter runs large reader surveys on how AI tools are changing software engineering work, e.g. the 2026 series 'AI's impact on software engineers' based on 900+ responses, and wrote 'The Software Engineer's Guidebook' (2023).

Founder of Code for America and former US Deputy CTO, she writes about why government struggles to put technology, now including AI, to work, and her newsletter hosts practitioner essays on AI as a push for process reform.

He is a widely read HR industry analyst who writes about how AI is changing HR work and workforce planning, and he sells an AI assistant for HR (Galileo) built on his firm's research.

His book 'Reshuffle: Who Wins When AI Restacks the Knowledge Economy' (2025) argues AI changes how work is coordinated and where value sits in a system, not only which tasks get automated.

His daily weblog documents hands-on use of LLMs and coding agents and popularized practitioner concepts such as 'prompt injection' and the risks of agent tool use.

His Understanding AI newsletter explains how AI systems work and checks claims about AI and jobs against data, e.g. his 2025 piece on Stanford evidence that AI is reducing hiring of young programmers.

Author of engineering-leadership books ('An Elegant Puzzle', 'Staff Engineer', 'The Engineering Executive's Primer', 'Crafting Engineering Strategy' 2025) and now writes first-hand about leading AI adoption inside an engineering org at Imprint.

Newsletter and podcast on product management; frequent guests on how product teams now work with AI.

Podcast where practitioners walk through exactly how they use AI in their work.

Newsletter and podcast for engineers building with AI.

Writes practical guides on evaluating AI systems (evals).

Writes on building and evaluating AI products, including how to work and compound with AI.

Writes first-hand accounts of working with coding agents.

Writes about malleable software and working with AI as a builder.

Daily newsletter on applying AI at work, heavy on agent and context practices.

Writes on agentic engineering in real codebases.

Writes about building with AI, including building your own benchmarks.

Marketer who writes daily about applying AI in marketing and analytics work.

Engineer who writes occasionally about how he uses AI in his own work.

Software pioneer writing about programming with AI assistants.

martinfowler.com, including Birgitta Böckeler's series on AI-assisted development.

Novelist and programmer writing carefully about his own AI use.

Designer and technologist who thinks in public about building with AI.

Writer and technologist on what AI changes about making software and running a firm.

Engineer writing frankly about AI and programming careers.

Engineering leadership writing, increasingly about managing teams around AI.

Opinionated writing on how companies work, including skepticism about AI claims.

Writes about AI's effect on the web and on technology workers.

Strategist writing about AI and mapping; posts rarely now.

Tracks how AI is changing legal practice.

Contract-drafting expert testing what AI gets right and wrong in legal drafting.

Writes on medicine, including AI in clinical practice.

Writing teacher on what AI does to learning to write.