Molly Kinder is leaving the Brookings Institution after three years running its generative AI and work project to start a new organization focused on what she calls the "messy middle" of the AI labor transition. Kinder's bet is that AI will not produce an overnight jobs apocalypse, but it will inflict concentrated, politically explosive losses on the highest-paid knowledge workers in the United States, and that the policy response so far is inadequate. She made the announcement in a podcast interview with Casey Newton at Platformer on June 9, 2026, alongside an essay laying out the framework.
Kinder's framing borrows a three-part schema developed with her Brookings colleague Palak Shah. Reality 1 is roughly the labor market as it exists today, with most jobs intact and limited measurable disruption from Claude, ChatGPT, and their peers. Reality 3 is the post-AGI world Silicon Valley promises, in which robots and AI can do nearly everything. Reality 2 — the messy middle — is the long period in between, where AI gets steadily more capable, partial automation hits specific roles hard, and most workers keep their jobs while a concentrated minority lose theirs.
The workers most at risk in that middle period, Kinder argues, are the ones who fared best in the last automation wave. Her test is blunt: if you can do your job locked in a closet with a computer, you are eventually going to be in trouble. That covers law, finance, consulting, sales, and a large clerical and back-office layer. It is, she notes, an inversion of the pandemic, when laptop workers were the safest cohort and frontline workers carried the exposure.
“If you can do your job locked in a closet with a computer, eventually you're probably going to be in trouble.”— Molly Kinder, Brookings Institution researcher
Key facts
- 01Molly Kinder is leaving the Brookings Institution after three years leading its generative AI and work project to start a new organization focused on the AI transition.
- 02Kinder frames AI labor disruption as a long 'messy middle' between today's intact labor market and Silicon Valley's promised post-AGI abundance.
- 03She argues white-collar knowledge jobs — law, finance, consulting, sales — will face displacement before blue-collar and service work.
- 04Kinder rejects universal basic income as a fix, calling instead for a workforce reinvestment fund, wage insurance, and white-collar apprenticeships.
- 05Her policy ideas would require companies cutting young workers to fund apprenticeships, and would back older displaced workers with wage insurance.
Kinder anchors the claim in OpenAI's own public dataset on task exposure to large language models, which scored every task in the economy for whether a tool like ChatGPT could meaningfully compress it. Aggregated up to jobs and sectors, the heaviest exposure sits in knowledge work tied to a bachelor's degree, plus the clerical layer underneath. Blue-collar, physical, and service-sector roles — restaurants, salons, repair shops — score low. Robotics will eventually change that calculus, but Kinder argues computer-based knowledge work is advancing faster.
She places the shift against 150 years of US labor market history. Agricultural work was mechanized, then manufacturing and blue-collar employment surged to upwards of a third of all jobs before declining. Since around 1980, knowledge and professional jobs in business, finance, accounting, and law have dominated job growth — what economists call skill-biased technological change. Computers, in that era, made the knowledge worker more productive and more in demand rather than replacing them.
Kinder's pointed example is the lawyer's office of the 1980s, where one attorney typically worked with one legal secretary handling typing, scheduling, and dictation. Computers absorbed the secretary's work and made the lawyer more valuable, not less. Large language models, she argues, could flip that pattern. If the tenth version of Claude in five years can actually lawyer rather than merely assist a lawyer, the cognition itself gets commoditized — and the workers who won the computer era become the ones who lose the next one.
“A world where most jobs are intact but there's a concentrated loss is still a world that is politically, societally, and economically explosive.”— Molly Kinder, Brookings Institution researcher
Her policy prescriptions deliberately reject the standard San Francisco answer of skipping straight to universal basic income. If displaced software engineers receive checks large enough to replace their salaries, she asks, why would anyone keep showing up to police streets, build houses, or staff hospitals? She wants targeted interventions instead: a workforce reinvestment fund that would require companies cutting young workers to pay for white-collar apprenticeships, wage insurance for older displaced workers, and, if good jobs grow genuinely scarce, a public effort to create new ones — effectively an industrial policy for knowledge workers.
Kinder is careful not to claim certainty about the timeline. She acknowledges no crystal ball, allows that white-collar impacts may land softer than she expects, and notes that current labor market data does not yet show large-scale disruption. Last week's Platformer interview with labor economist Kathryn Anne Edwards, which AI Chat Daily covered alongside Microsoft's Mustafa Suleyman walking back his own white-collar automation claim, struck a similar note of measured concern about the weak US safety net without predicting a permanent idle class.
“You've just destroyed the labor market.”— Molly Kinder, Brookings Institution researcher
The skeptic's case is straightforward: aggregate unemployment remains low, no economy-wide dislocation has shown up in the data, and previous predictions of technology-driven mass unemployment have repeatedly underestimated job creation in adjacent categories. Kinder's own framing concedes the point — most jobs survive in her scenario — but argues the political and economic consequences of concentrated losses among lawyers, consultants, and early-career professionals would be severe regardless of the aggregate number.
Kinder's departure from Brookings to build a dedicated organization is itself a market signal. The think-tank-to-operator move suggests that the people closest to the AI labor data believe the policy infrastructure is not being built fast enough at existing institutions. For AI companies including Anthropic, OpenAI, and their enterprise customers, the practical implication is that the political backlash to white-collar deployment may arrive before the productivity gains fully compound — and the cohort organizing the response is now well-funded, credentialed, and explicitly opposed to the UBI-shaped answer the industry has often defaulted to.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
