In late March, I started receiving daily texts from the federal government about AI. “🇺🇸AI is changing how we work and live,” one message read. “You might feel curious, skeptical, or unsure—that’s normal.” I had enrolled in an AI-literacy course from the Labor Department created to help workers succeed in the ChatGPT economy. The weeklong program, created in partnership with an AI start-up and delivered by text message, was supposed to equip Americans with “foundational AI skills,” according to an agency press release. But the government’s texts were not reassuring. One message encouraged me to ask a bot for “side hustle ideas.” Another suggested that I brush up on my AI skills by doodling: “Grab a friend and see whose drawing of a hippo AI can recognize … and whose it mistakes for a lumpy potato. 🥔.”When it comes to AI’s threat to jobs, America has acted like a deer in headlights. For all the catastrophic messaging about looming layoffs, efforts to prepare for a major transition have lagged. After speaking with a roster of the nation’s top business and political leaders, my colleague Josh Tyrangiel concluded earlier this year that no one seemed to have a plan for how to proceed. Government efforts to shore up the labor market have been largely underwhelming. AI-literacy courses like the one I took are hardly a sufficient response if you believe that mass job loss is forthcoming; the Trump administration announced a goal last year to expand apprenticeship programs, but progress on that front has been middling. Meanwhile, the tech industry has barreled ahead in its efforts to develop more capable models. Earlier this year, Anthropic’s Dario Amodei warned that “AI isn’t a substitute for specific human jobs but rather a general labor substitute for humans.”But over the past several weeks, politicians and tech executives have started to make a show of more seriously preparing the country for a future of AI layoffs. As promising as these new efforts are, many of them also end up serving the interests of Silicon Valley.Consider a flashy proposal for state ownership of AI companies. Last month, Senator Bernie Sanders proposed the AI Sovereign Wealth Fund Act, which would give the federal government a 50 percent stake in major AI companies including Anthropic and OpenAI. The act would also give an appointed commission voting shares to help “stop bad decisions that will reap massive job loss,” Sanders said on a press call. The legislation, which calls for the tech industry to hand over trillions of dollars worth of stock, is unsurprising coming from Sanders, who regularly takes aim at “Big Tech oligarchs.” What’s surprising is the fact that at least one of the alleged Big Tech oligarchs is on board with a watered-down version of Sanders’s proposal.After Sanders first announced his plan, Sam Altman requested to meet with the senator. The OpenAI CEO reportedly told Sanders that although he does not endorse the exact plan, he supports the idea of public ownership of AI companies. Altman has been flirting with such ideas since even before ChatGPT was launched, and in April, OpenAI endorsed the creation of a “public wealth fund” that would give every citizen “a stake in AI-driven economic growth.” More recently, according to the Financial Times, Altman has had recent conversations with the Trump administration about public ownership of OpenAI and its competitors; the company is reportedly floating giving up a 5 percent stake.[Read: Why everyone is suddenly talking about ‘universal basic capital’]Still, it’s difficult to imagine a future in which our dysfunctional Congress passes a law to take a stake in AI companies, and then the government seamlessly redistributes the earnings to the rest of us. And regardless, such arrangements are unlikely to save a flailing labor market. Sanders’s aggressive wealth fund, for example, aims to start by paying roughly $1,000 to each American annually—which is not trivial, but is also far from a solution to mass unemployment. And private companies don’t give away equity out of altruism. If the government were to seize control of AI companies, the country could end up responsible for bailouts (although Sanders’s bill prohibits use of the fund for such purposes), and public ownership could dilute the demand for more onerous regulation. Perhaps that’s why Gwynne Shotwell, the president of SpaceX, home to Elon Musk’s AI efforts, announced this week that she and her husband would give stock to Trump Accounts, the savings accounts for children that just launched. Anthropic has also said that it would like to invest in researching “AI sovereign wealth funds” as a possible response to extreme unemployment.AI companies are eager to be seen as supporting other initiatives too. In May, OpenAI’s nonprofit foundation, which owns roughly a quarter of the AI company, announced that it would spend $250 million on grants and partnerships to support the forthcoming economic transition, while also building its own team to seed new ideas. And last month, Anthropic committed $350 million toward similar efforts, much of which will fund “major research trials and program evaluation on promising public policies.” Specifics are still being worked out, but both organizations have noted an interest in, for example, investing in efforts to better measure how AI is affecting labor markets, as well as unemployment-insurance reform. (Anthropic declined to comment; a representative for the OpenAI Foundation pointed me to the company’s April policy paper.)Economists, for their part, remain unsure about what types of policy interventions could best mitigate future problems, or really how big those problems are likely to be. AI has not yet led to widespread job losses, and many economists aren’t convinced that a jobs apocalypse is inevitable. But even absent the extreme, if moderate labor displacement occurs quickly, America could be in for a real shock. In that light, one new nonprofit, Raise Us, launched last month with the intention of testing out different policies to see what works. Led by Gina Raimondo, Joe Biden’s commerce secretary, and Eric Holcomb, a former Republican governor of Indiana, the organization has already raised more than $500 million from private companies and philanthropies (including the Rockefeller Foundation; Melinda French Gates’s group, Pivotal; and Emerson Collective, which is the majority owner of The Atlantic). Some of that funding comes straight from Silicon Valley: Anthropic, Amazon, Microsoft, and the OpenAI Foundation have contributed cash to the effort and are among the organization’s founding partners.Raimondo, who worked extensively on AI policy during the Biden administration, told me that she hopes that by doing smaller-scale experiments now at the state level, the organization will gather data on which interventions are most effective, helping inform federal policy efforts down the line. (A set of both red and blue states—Utah, Connecticut, Arkansas, and Maryland—have signed on as initial government partners.) For example, the group is interested in testing wage insurance. Under such a policy, if someone loses a job and accepts work at lower pay, wage insurance would aim to temporarily fill some of the gap. There is “some positive evidence” in favor of the policy, David Autor, an MIT economist who serves as an unpaid adviser to Raise Us, told me, but it has “never been done at any significant scale in the United States.” The group also plans to invest in a slew of other interventions, including apprenticeship programs and career-navigation tools. Overall, the group’s efforts will likely be useful in generating real-world policy evidence that is “undersupplied” and “extremely valuable,” Susan Athey, a Stanford economist who is not affiliated with Raise Us, told me.[Read: The AI backlash could get very ugly]At the same time, the tech industry’s association with Raise Us is also a play to boost its flagging reputation. Brad Smith, the vice chair and president of Microsoft, told me that the Raise Us coalition is similar to the public-private partnerships that formed in response to the pandemic. It’s a compelling comparison, but a flawed one: Imagine if the vaccine companies involved with Operation Warp Speed were also engineering more sophisticated viruses. Silicon Valley has wedged itself into a strange position. AI companies are racing to automate work while simultaneously preparing to parachute in as saviors.Of course, shepherding a smooth labor transition is in Silicon Valley’s self-interest. An AI backlash is brewing, and tech companies’ reputations will only further degrade if layoffs mount. The industry also has reason to worry about the federal government taking far more dramatic action against it. But all this hullabaloo around wealth funds and public-private partnerships is about much more than just optics. AI companies are leveraging the current anxiety over layoffs to induce greater dependence on their bots. Several industry-supported efforts to ward off a jobs crisis involve training people to better use AI. In Maryland, for example, Raise Us is supporting the launch of a start-up-accelerator program, which will provide displaced workers with financial assistance—and access to AI tools to help them start businesses. And almost half of Anthropic’s $350 million commitment will go toward creating a national fellowship called Claude Corps. Through the program, Anthropic will pay early-career workers $85,000 each to spend a year working to integrate Claude across at least 400 nonprofits. The solution to Claude’s threatening entry-level workers’ jobs: Hire young people to further evangelize Claude.