By Lisa Bonos, Stanford Graduate School of Business Image: Galina Nelyubova – Unsplash How will advances in artificial intelligence change jobs, companies, and the economy at large? What can leaders and employees do to prepare for a world that will be reshaped by AI? Erik Brynjolfsson — a professor (by courtesy) of economics and of operations, information, and technology at Stanford Graduate School of Business and director of the Stanford Digital Economy Lab — has spent years researching these questions and anticipating the transformative effects of new technologies. Among his many recent projects, he’s been tracking how companies are deploying AI, identifying which workers can benefit most from AI, and calling on economists, AI researchers, and policymakers to act now “to build the incentives, guardrails, and institutions needed to steer AI in a direction that complements humans and benefits society. ” Brynjolfsson spoke with Insights about what it will take to successfully reimagine companies and institutions so that humans can work collaboratively with AI to increase their productivity. Based on your research, how can AI be deployed while also addressing its impacts on real people? Erik Brynjolfsson: At the Digital Economy Lab, Elisa Pereira, Alvin Wang Graylin and I studied 51 AI deployments that delivered value in a variety of industries, including manufacturing, financial services, and technology. We found that 77% of the hardest challenges weren’t technical but rather were change management, data quality, and process redesign. The key to success was almost never the AI model. It was the organization. The same technology led to several different answers. Headcount reduction was important, but it was only 45% of cases. The rest were instances where workers were redeployed to higher-value tasks. How do you get the better answer? Think tasks, not jobs. Almost no job is fully automatable, while almost every job has tasks that AI can change. Once you deconstruct work that way, displaced capacity shows up as the ability to redeploy workers instead of cutting headcount. One of your indicators, the Canaries Dashboard, shows the starkest employment decline for 22-to-25-year-olds in AI-exposed occupations compared with workers of older generations. How do you see entry-level roles changing? Erik Brynjolfsson: In a recent study with Bharat Chandar and Ruyu Chen, we found a 19% employment decline for 22-to 25-year-olds in the most AI-exposed occupations relative to their peers in less exposed occupations. But the more interesting finding is that the decline was greatest in occupations where AI automates rather than complements human labor. The routine production part of junior work — the first draft, the basic query, the initial screen — is what these systems do well. What’s left is judgment, verification, client contact, and orchestrating AI. So for many, the early-career on-ramp gets steeper. What advice do you give students heading into the workforce? Erik Brynjolfsson: My advice is to build with people unlike you. And optimize for making something work. This past spring, I taught an economics and computer science class called The AI Awakening. It was about half GSB students and half engineers and computer scientists, with guests from leading AI labs, venture capital firms, and cutting-edge startups. Teams of three had 10 weeks to ship not just a business plan but also a working agentic prototype. The students learned from each other, especially when they were from different schools, and built something real. If AI fluency isn’t enough to protect your job, what will? Erik Brynjolfsson: Fluency is important but it’s not enough. Being AI-fluent in 2026 is like being email-fluent in 2005. What protects you is learning how to ask the right questions to guide your work with AI. Another key is having accountability for the outcome. That builds relationships and trust. Those are the tasks where the machine makes you more valuable instead of replacing you. The practical move is to assess: Which of your tasks does AI substitute for? Which does it amplify? Then shift your weight toward the second list before someone shifts it for you. That’s not a skill you acquire. It’s a habit of reallocating your time. How should lawmakers and technologists prepare for the seismic shifts that AI will bring about? Erik Brynjolfsson: I recently co-organized a statement about AI and transformation with fellow technology-minded economists Ajay Agrawal, Anton Korinek, and Tom Cunningham. It was striking how broadly so many economists, who often disagree on other topics, agreed about the impact of AI. Measure the impacts first. Second, stop forecasting and start building. Institutions take years to construct, and transformative AI may not wait for them. Third, make the complementary path the profitable one. Tax neutrality between labor and capital and collaborations that reward human-plus-machine performance. What do the majority of workers who barely use AI understand that the builders don’t? Erik Brynjolfsson: Pew finds that about 65% of American workers say they use AI little or not at all. When new general-purpose technologies like AI first emerge, it can take years of intangible investment before measured gains materialize. My collaborators and I dub this dynamic the productivity J-curve. Now we’re at the point in the J-curve where early investments have not yet yielded large advances, so even if most workers aren’t using AI yet, they will be soon. Our research on how AI is deployed makes it concrete. One executive told us all the hard work is in process documentation and data architecture. Only 6% of the firms we studied had clean data to start from. The people doing the work knew that already. In 2036, what indicator would you consult to see whether we got this right? Erik Brynjolfsson: GDP-B, a new measure of welfare and growth I devised with Stanford colleagues that accounts for how much consumers benefit from goods and services, not just how much they pay for them. Our assessment puts U. S. consumer surplus from generative AI at roughly $172 billion this year. Essentially, none of it shows up in traditional measures of GDP. GDP was one of the great inventions of the 20th century, but it misses an increasing share of value in the 21st century. We need a new metric like GDP-B. This piece was originally published by S tanford Graduate School of Business. Fact-Checked by Irfan Ahmad. Read next: • Researchers Explore Whether AI Could Become Conscious In The Future • Advertising is coming to AI chatbots — and it could influence the answers you get • FBI Warns of AI-Assisted Scams Impersonating Government Officials as AI Fraud Shows Higher Profitability
“Almost Every Job Has Tasks That AI Can Change”. Economist Erik Brynjolfsson says AI’s transformation of work is just beginning and our institutions aren’t ready for it
RELATED ARTICLES
