Harvard economist Jason Furman told Axios that if he returned today to his old role as chair of the White House Council of Economic Advisers (CEA), he would rely on AI tools for much of the council’s day-to-day analytical work instead of staff economists. The remark sparked debate online, particularly among the loose-knit community formerly known as Econ Twitter.
Concerns from experienced policy practitioners
Several veterans of economic policy worry that an overreliance on AI by future policymakers could carry real risks, or at least that Furman understates the importance of human economists in advising leaders.
Jared Bernstein, who chaired the CEA under President Joe Biden, wrote on Substack that CEA work is "highly bespoke," because advisers must explain data flows and many other details to very busy officials who want more than market-focused analysis (for example, on CPI, jobs, GDP). Bernstein said leaders want to know how a report fits into the administration’s narrative, agenda, or the president’s schedule, and that inside the White House "the human skill of reading-the-room is paramount." He added that, while not an AI expert, his sense is AI is poor at that skill and probably will remain so.
Claudia Sahm, a former Federal Reserve economist, raised a common concern from senior people in consulting and investment banking: if junior positions disappear, where will future senior leaders come from? Sahm noted that Furman’s former staff "went on to found the Yale Budget Lab, run Treasury economic policy and a Fed institute, serve in later [administrations], and publish research on trade, environment, and taxes. That's what '20 staff' become." She wrote, "No model does that."
Anna Wong, Bloomberg’s chief U.S. economist, warned that the ability for a senior person to "not depend on anyone" can be very empowering — with just an idea, a senior person could go far. She also raised a speculative concern that over time AI systems themselves, through training, might become primary idea generators, a development she described as "a disaster waiting to happen" that social policy needs to address.
Alternative view: AI can increase demand for labor
Casey Mulligan, who served on the CEA in President Donald Trump’s first term and now is chief economist at the Department of Health and Human Services, argued that AI’s effects on policymaking labor are not obvious. Mulligan said he built AI tools in two federal offices and found that because research could be done faster and better, the offices needed more economists, not fewer, due to increased demand for their work. He added that the same pattern appears in the small business sector. This idea — that labor-saving technologies can increase demand for labor — reflects the Jevons Paradox.
Bottom line: design organizations where AI and people complement each other
Stanford economist Neale Mahoney urged that the appropriate response is to be the kind of labor-augmenting technological change one wants to see. He said it is up to older generations to build organizations where people and AI complement each other and to prepare younger generations for a world with ever-improving AI.
In sum, the conversation is not binary: proponents highlight efficiency gains from AI, while critics stress risks to talent development, contextual judgment, and broader social consequences. Many commentators converge on the need for deliberate organizational design and training to ensure AI augments — rather than displaces — future policy expertise.
Key takeaways
- Jason Furman said he would favor AI tools for much of the CEA’s analytical work.
- Critics including Jared Bernstein and Claudia Sahm warned this could weaken future talent pipelines and lacks human "room-reading."
- Others, like Casey Mulligan, argue AI may actually raise demand for economists by increasing productivity.
- Neale Mahoney called for building organizations that let people and AI complement each other and for preparing young professionals for an AI-rich future.
