In early May, Sam Altman traveled to Washington to tell a story about America. Appearing before a Senate committee, Altman described how he came of age as the internet took off, how he stayed up late in his family’s attic and learned to code on products that were invented in the United States—a personal computer, its silicon chips and accompanying software. That early experience with the “spirit of American innovation,” Altman told the senators, put him on a path to found OpenAI, launch ChatGPT, and set off the AI boom. “I think America is just an incredible and special thing,” he said, “and it will not only be the place where the AI revolution happens but all the revolutions after.”
Altman’s written testimony, which was submitted to the Senate, added an important asterisk that he did not speak aloud that day. “This future can be almost unimaginably bright,” OpenAI’s CEO wrote, but only if “an American-led version of AI, built on democratic values like freedom and transparency, prevails over an authoritarian one.”
Silicon Valley’s tech giants have recast the AI boom not just as a matter of scientific and economic advancement but as a clash of civilizations. They are fixated on competition with China, and the idea that Chinese AI, should it outpace its American counterpart, will extend a repressive surveillance state into the rest of the world. Despite the rhetoric, however, it is not at all clear that AI companies are doing anything themselves to uphold American freedom. In fact, they seem much more interested in what America can do for them.
China is not a new worry. For many years, Eric Schmidt, Sheryl Sandberg, and other tech leaders have warned about the need to outpace China in various technologies—including AI, quantum computing, and 5G—but these concerns have become louder since the launch of ChatGPT. Dario Amodei, the CEO of OpenAI’s rival Anthropic, wrote last fall that “AI-powered authoritarianism seems too terrible to contemplate.” Democracies, he continued, must band together to stay ahead. Certain American ideals have always been embedded in Silicon Valley—individualism, the belief that you can build anything you can dream of, the promise that hard work will pay off. But where the previous tech revolution—the arrival of the social web and smartphones—was described by tech titans as a path toward global connection and democratization worldwide, the arrival of artificial intelligence has taken on a decidedly “America First” flavor.
There are two main ways to understand what is happening here. One is the idea that the world’s most powerful tech companies are bending toward the current ruling political class in America, casting their own missions in terms that they believe will be favorable to MAGA Republicans, generally, and Donald Trump, specifically, to avoid regulation and further enrich themselves. The other way to understand this moment is to see that the world’s most powerful tech companies are now producing dramatic innovation that, in an earlier era, would be solely under the purview of the U.S. government, and they recognize the responsibility they have to the world’s citizens. The truth of what’s happening is somewhere in the middle.
Since Trump’s victory, Altman, Amodei, and their cohort have become more forceful in their warnings: Generative AI could supercharge China’s economy, propaganda apparatus, espionage capabilities, and coercive tactics—and even make its military more powerful than our own. Last month, Anthropic published a report on how to “Build AI in America” and “compete with China.” Alexandr Wang, the former CEO of Scale AI and now the chief AI officer at Meta, has told the Trump administration that the U.S. and China are in an “AI war.” Earlier this summer, executives from OpenAI, Meta, and Palantir joined a new Army detachment, the Executive Innovation Corps, to serve as part-time advisers. The group is tasked with “marrying the nation’s most innovative private companies with our most important military missions,” Palantir’s chief technology officer, Shyam Sankar, wrote in an article for The Free Press about joining the unit.
In the U.S., democracy and AI have become conjoined in a sort of catchphrase. The highest echelon of American AI firms—including OpenAI, Anthropic, Google, Microsoft, Scale, Meta, and Palantir—all made recommendations to the Trump administration as it developed its “AI Action Plan,” appealing to variations on “democratic AI,” “American” ideals, and “western” values. Similar invocations of democracy and American leadership litter the lobbying documents, policy proposals, interviews, and congressional testimonies of these companies and their executives. And in the broadest terms, these companies got exactly what they asked for when the AI Action Plan was unveiled last month: Trump, having already cut off a number of Biden-era AI regulations before they could go into full swing, gave the industry a green light to develop new products with limited oversight, and a promise that few, if any, roadblocks would stand between them and the buildout of economically and environmentally costly infrastructure. “Our children will not live in a planet controlled by the algorithms” of our “adversaries,” President Donald Trump declared when he announced the plan. Of course, the algorithms and technology shaping our lives are not developed or closely overseen by the U.S. federal government, either. Pentagon laboratories are not spurring the AI boom, which purportedly holds the keys to America’s military and economic future. Rather, the private sector is.
But “democratic AI,” ascendant under a president who has already tried to overturn an election, may prove to be a misnomer. Tech firms have long demonstrated a willingness to do what is best for them, not the greater good—and now the AI boom has offered a way to conscript the American project to advance their techno-utopian visions, not the other way around. Altman may be speaking honestly when he invokes the great tradition of American innovation, but that does not change the fact that aligning the political apparatus behind OpenAI grants him obvious business advantages. OpenAI recently launched OpenAI for Countries, its version of the Marshall Plan: a project to “spread democratic AI” to other nations in the form of ChatGPT. (OpenAI and The Atlantic’s business division have a corporate partnership.) Microsoft, Amazon, Nvidia, and other prominent AI firms, too, have described spreading their products and services around the globe as a sort of diplomacy on the behalf of the U.S. Meanwhile, OpenAI and its rivals have also become more aggressive in their efforts to keep individual users walled into their chatbots, where the companies can accrue valuable data and charge subscription fees. Democratic AI, similarly, can feel like a user-acquisition strategy.
In recent weeks, OpenAI and Anthropic have announced that they are providing federal agencies a year of premier subscriptions to ChatGPT and Claude for only $1, covering all of their employees; Google, not to be outdone, is offering a Gemini subscription for 50 cents per agency. “Federal workers deserve access to the most capable AI tools to better serve the American people,” Anthropic said of its offer, which also extends to the legislative and judiciary branches, as if not using AI disservices the nation. Once the trial period ends, these programs could cost the full price of an enterprise contract, putting the federal government on the hook for millions of dollars, if not far more. Democracy has been recast as a project for tech firms to advance; they’re not software developers so much as they are regime builders. All of this is a strong echo of the rise of social-media giants, digital empires that have influenced the outcomes of elections and stoked some of the worst human impulses imaginable. Hillary Clinton once described interacting with Mark Zuckerberg as “negotiating with a foreign power” to my colleague Adrienne LaFrance.
American politicians have been wholly unsuccessful in reining in that power, and now, as the AI boom brings Silicon Valley’s ambitions to new heights, they are positioned more than ever as industry cheerleaders. Seen one way, this is classic conservatism: the championing of America’s business rulers based on the belief that their success will redound on the nation. Seen another, it is a dereliction of duty: elected officials willingly outsourcing their stewardship of the national interest to a tiny group of billionaires who believe they know what’s best for humanity.
The tech industry’s new ambitions—using AI to reshape not just work, school, and social life but perhaps even governance itself—do have a major vulnerability: The AI patriots desperately need the president’s approval. Chatbots rely on enormous data centers and the associated energy infrastructure that depend on the government to permit and expedite major construction projects; AI products, which are still fallible and have yet to show a clear path to profits, are in need of every bit of grandiose marketing—and all the potentially lucrative government and military contracts—available. Shortly after the inauguration, Zuckerberg, who is also aggressively pursuing AI development, said in a Meta earnings call, “We now have a U.S. administration that is proud of our leading companies, prioritizes American technology winning, and that will defend our values and interests abroad.” Altman, once a vocal opponent of Trump, has written that he now believes that Trump “will be incredible for the country in many ways!”
That dependence has led to a kind of cognitive dissonance. In this still-early stage of the AI boom, Silicon Valley, for all its impunity, has chosen not to voice robust ideas about democracy that differ substantively from the whims of a mercurial White House. As millions of everyday citizens, current and former government officials, lawyers and academics, and dissidents from dictatorships around the world have warned that the Trump administration is eroding American democracy, AI companies have remained mostly supportive or silent despite their own bombastic rhetoric about protecting democracy. There have been some instances of pushback, although only in reference to hyperspecific tech policies: Microsoft recently said it could sue if the Trump administration, or any government, were to force it to suspend cloud operations in Europe as part of a trade war, and Amodei, the CEO of Anthropic, published an editorial in The New York Times criticizing the Republican proposal to outlaw state-level AI regulation. Anthropic has also spoken against the Trump administration’s lifting of some restrictions on AI-chip exports to China, on the grounds that export controls are necessary to slow the development of Chinese AI models.
Trump, meanwhile, has targeted protests with military force, retaliated against political foes, openly defied the courts, mobilized the military against lawful protestors, used his office to enrich himself and his family, and signed an executive order against “woke” AI, attempting to narrow Americans’ access to material that chatbots would otherwise present. These actions would all seem to fit the very mold of the “authoritarian” power the AI companies claim to fear, though they have not spoken out against them.
I reached out to OpenAI, Anthropic, Google, Microsoft, Scale, Meta, and Palantir, seeking interviews to discuss these issues. When asked or upon receiving no response, I followed up with a detailed set of questions. Meta acknowledged but did not respond to my questions, and Google did not respond at all. Liz Bourgeois, a spokesperson for OpenAI, wrote, “We believe the United States is the best place in the world to build democratic AI,” adding “there’s broad agreement across both Republicans and Democrats that the U.S. should lead in this effort, and we believe the importance of democratic AI transcends partisan politics.”
Spokespeople for Scale, Microsoft, and Anthropic declined to comment but pointed me to recent company statements, including congressional testimonies from their executives that described a belief that advancing AI built in America is necessary to strengthen democratic values and combat China. “AI built in authoritarian nations will—no matter what the personal preferences are of the people in those countries building it—be inescapably intertwined and imbued with authoritarianism,” Jack Clark, Anthropic’s chief policy officer, recently told Congress. Courtney Bowman, the global director of privacy and civil-liberties engineering at Palantir, told me over email that helping the government run smoothly is important “regardless of the administration in power, even in times of deep disagreement about the direction of our democracy itself.” The American project can persist, Bowman wrote, only if federal agencies such as the military, CDC, and IRS—all of which Palantir has contracted with—persist as well.
Trump has restaffed the military with loyalists, slashed the IRS’s workforce, and filled the CDC with science denialists. The AI industry’s belief that its products can drastically improve the world may be genuine, but the rationale underlying its insistence is shallow at best. AI’s titans seem far less concerned with America as a democratic project and more interested in it as a brand, financer, and regulatory backer. Perhaps AI executives are also now AI patriots, but that is a secondary occupation; faith in technological acceleration, and maintaining their own power, still transcends a commitment to democracy.
The post Do AI Companies Actually Care About America? appeared first on The Atlantic.
