AI & Tech Brief: AI’s Economic Reckoning
- More than 200 economists and AI luminaries are urging action to address social and economic AI disruption. Daron Acemoglu, an MIT economist and noted AI productivity skeptic, tells me 10 percent job displacement within the next decade is likely.
- The House Democratic Caucus is holding a listening session on AI policy Tuesday as the push for a data center moratorium is going mainstream in the party.
- OpenAI and Anthropic employees (among them many soon-to-be billionaires and decamillionaires) are exploring philanthropic opportunities to spend all that money.
This is Benjamin Guggenheim. Welcome to WP Intelligence’s AI & Tech Brief, where we examine the transformative technology of artificial intelligence at the intersection of innovation, policy and power.
“AI may become radically more powerful over the next 10 years,” says the statement released Monday. “This could drive an unprecedented transformation of our economy, larger than the Industrial Revolution, but unfolding over a vastly shorter time frame.”
The letter counts some notable signatories hailing from the social sciences, including economists Joseph Stiglitz and Jason Furman, as well as former Federal Reserve chairman Ben Bernanke. There are also big names in tech such as former Google CEO Eric Schmidt and Jack Clark, co-founder of Anthropic.
“I think the statement is too weak because it doesn’t establish any principle that the government should seek,” Mordecai Kurz, an economist at Stanford University, tells me. (Kurz nonetheless has signed the letter.)
Kurz says that action should be oriented around the idea that companies have a choice of how to design AI systems. Tech should be designed specifically to augment human workers, not replace them, he says.
I also caught up with Daron Acemoglu, a Nobel laureate signatory and economist at the Massachusetts Institute of Technology, about his thoughts. Acemoglu has previously criticized much of the AI discourse as “brainless,” and has expressed deep skepticism of the AI productivity gains that Wall Street has forecasted.
Acemoglu tells me that he has indeed been taken by surprise by some of the extraordinary acceleration of AI capabilities over the past year.
“What we saw at the end of ‘25 and the beginning of ‘26 was an acceleration. I think the agentic AI models are certainly much better in terms of a number of tasks, such as coding and other sort of simple cognitive tasks,” says Acemoglu, adding that he’s still skeptical of the projections by some of the prominent AI luminaries about entry-level jobs.
Take “[Anthropic CEO] Dario Amodei’s claim that 40 percent, 50 percent of all entry-level positions will be done by agentic AI within a couple of years. I think that’s still unrealistic,” he said. “If you mean that 10 percent of tasks, 10 percent of jobs, can be done by AI within a 10-year horizon, definitely. That’s a lot. … The disruptive effects of that would be huge.”
The Legislative Chessboard
Meanwhile, the legislative body charged with crafting a coherent approach to the disruptive effects of AI is still getting its act together.
According to a memo I obtained exclusively, House Democrats are holding a listening session Tuesday afternoon at 3 p.m. on AI policy. The session for lawmakers is hosted by the House Democratic Commission on AI, which was created by House Democratic Leader Hakeem Jeffries (D-New York) and is co-chaired by Reps. Josh Gottheimer (D-New Jersey), Ted Lieu (D-California), and Valerie Foushee (D.-North Carolina).
“The big three thing [on the agenda]: jobs, energy and child safety,” said a senior Democratic aide who works on AI policy and spoke on the condition of anonymity to speak candidly.
It comes as the House Democratic Caucus is reckoning with polarization within its ranks about how to move forward on AI policy. Once a policy fixture of far-left progressives, the push to implement a moratorium on data centers has become more mainstream — with Rep. Frank Pallone Jr. (D-New Jersey), the highest-ranking Democrat with jurisdiction over such energy issues on the Energy and Commerce Committee, calling for a freeze on data center construction at a hearing in June.
The stance is directly at odds with moderates in the party concerned about the AI race with China, such as Gottheimer, who has said on X that “we need to keep building the infrastructure that allows us to lead.”
Apparently referencing the influence of pro-industry AI super PACs in Congress, Rep. Greg Casar (D-Texas), the chair of the Congressional Progressive Caucus, told Politico that Democrats are “in the wilderness” on AI policy and that “we absolutely cannot let the AI money silence us.”
Many senior Democrats also privately say it won’t be advantageous to put forward legislative text on a frontier AI framework until the next Congress — when the party expects to take back the House.
“Why negotiate with yourselves when you don’t know what the 120th Congress will look like?” the senior Democratic aide said.
The Senate side: In the upper chamber, talks between Sen. Marsha Blackburn (R-Tennessee) and the White House appear to have sputtered ever since the House passed a suite of children’s safety bills at the end of June — which run contrary to the children‘s online safety priorities of Blackburn and a bipartisan group of lawmakers in the chamber.
Blackburn and others had hoped to pair a different set of online safety legislation with a narrow “preemption” of state AI laws long coveted by the AI industry.
Demis steps in: Demis Hassabis, the CEO of Google DeepMind, posted a rare essay on X on Tuesday arguing that humanity is at the “foothills of the singularity” of “artificial general inteligence,”or when AI can surpass human cognitive capabilities across all tasks. He writes that the impact of the technology would perhaps be 10 times that of the Industrial Revolution.
Hassabis calls for a private-public partnership with a board comprising technical experts to regulate AI — with significant funding to attract “world-class technical talent. Hassabis also says that testing frontier AI models could at first be voluntary but then subsequently be formalized, such that models need to pass the vetting in order to be released to the U.S. market.
It’s an idea that lawmakers such as Rep. Sam Liccardo (D-Calif), who represents Silicon Valley, have alluded to before in light of lagging progress on congressional AI policy. And a proposal for an independent regulatory body could become increasingly appealing to large industry players that are desperate for clear regulatory guidelines — as opposed to the ad hoc approach the administration took with Anthropic’s Fable model.
The AI industry sees the administration’s June executive order for a voluntary framework as helpful. But there still needs to be way more meat on the bones of a substantive regulatory regime.
“I think this is something that is worthy of debate and should be a real focus. … Because the world’s not going to wait, and there’s a fair amount of discontent with the uncertainty right now,” said Paul Lekas, executive vice president for global public policy & Ggovernment affairs of the Software Information Industry Association.
“We don’t have any established process, system, framework [or] organization to do assessments and evaluate models before they’re released to help identify potential risks that can be mitigated,” Lekas said.
Philanthropy x AI
As both OpenAI and Anthropic ready initial public offerings, soon-to-be billionaires and centimillionaires are starting to explore philanthropic options for how to spend all that money, Nitasha Tiku reports for The Post.
Nitasha reports that, last Thursday evening, employees from Anthropic and the OpenAI Foundation, the nonprofit parent entity of OpenAI, attended an event at the American Bookbinders Museum in San Francisco titled, “What should we do with all this money?”
The event was hosted by a magazine funded by Coefficient Giving, a philanthropic group founded by Facebook co-founder Dustin Moskovitz and his wife, Cari Tuna, with roots in the do-gooder effective altruist movement.
For context: Anthropic’s seven co-founders, who are all in line to be billionaires, have pledged to donate 80 percent of their wealth — and many of the company’s staffers have signed up for a generous donation-matching program.
Meanwhile, the OpenAI Foundation holds 26 percent of the for-profit company, which is valued at more than $180 billion.
What happened: Adherents of the effective altruist movement have supported different philanthropic issues ranging from efforts to alleviate poverty to supporting animal welfare. But Nitasha reports that the evening featured a lively debate about how best to leverage newly formed AI wealth.
“We’re talking about Open AI and Anthropic; those are two AI companies that have very distinctive views about the future of the world based around AI,” Coefficient Giving CEO and co-founder Alexander Berger told the crowd.
“People who are thinking about AI safety, AI for good … should expect world historical increases in the speed and shape of that giving.”
The upshot: Expect that newly wealthy employees from OpenAI and Anthropic will be exploring ways to get involved in AI policy and regulation in Washington, D.C., too.
Already, new groups such as Control AI, which is warning about the threats of superintelligence (when AI outperforms human minds), are establishing new presences in the capital and exposing policymakers to ideas that would have seemed outlandish just a year ago.