Former President Barack Obama used a campus appearance in upstate New York to demand federal regulation of artificial intelligence, dismissing the Trump administration's market-driven strategy as reckless at a moment when even AI builders warn the technology could pose existential risks.
Obama spoke Friday at Colgate University in Hamilton, New York, in a sit-down interview with university president Brian Casey before an auditorium of students. He framed AI as a powerful but dangerous tool that demands government oversight, and he took direct aim at an unnamed Trump adviser's argument that market forces and the threat of lawsuits would keep the industry in check, Breitbart News reported.
The former president's remarks landed in the middle of a fast-moving week for the AI industry, one that included alarming safety disclosures from OpenAI, a public call from Anthropic's CEO to slow the pace of frontier development, and a researcher's resignation over fears that the companies building AI "earnestly believe that it could kill us all by the end of the decade."
Obama told the Colgate audience that AI should be understood as a machine, not a sentient being, but warned that getting the technology wrong carries severe consequences.
"Really useful stuff, but if we get this wrong, it could be chaotic and potentially catastrophic."
He then turned to the Trump administration's posture. Without naming the specific adviser, Obama described hearing a senior Trump figure argue that companies would solve AI safety on their own because of financial liability and the risk of lawsuits.
Obama rejected that logic outright, comparing it to a hands-off approach the country abandoned generations ago in other high-stakes industries:
"That's not how we treat airlines, or drug companies, or food companies. We now have 150 years of experience just letting the market do whatever it wants when it comes to things that deeply affect our health and our safety."
He went further, characterizing the current White House as both unwilling and unable to build the regulatory framework he says the moment requires:
"And so there's no replacement for an effective government regulatory structure. And the challenge we have is that obviously, we have a president currently that has said that's for losers. So, we've got a gap, at least of a couple of years, where things are moving really fast, and we have an administration that is not capable of or willing to, I would argue both, of putting together a serious regulatory framework, which is why I actually think we should be encouraging some sort of voluntary industry restraint, not because it's a replacement, but because that's probably the best we can do for now until we can get Congress and the White House to start getting serious about this."
That framing, voluntary industry restraint as a stopgap because Washington won't act, is a familiar Obama move: position yourself as the reasonable adult while casting the sitting president as negligent. But it glosses over a basic tension. Obama held the White House for eight years and left office without establishing the federal AI regulatory structure he now demands. The technology was less advanced then, but the call for government oversight is easier to make from a university stage than from behind the Resolute Desk, where trade-offs are real and bureaucracies are slow.
It is worth noting that Obama has a pattern of working behind the scenes on political matters while choosing his public moments carefully. His AI remarks are no exception, timed to a week of industry turmoil that gave his argument maximum wind.
This is not the first time Obama has waded into the AI regulation debate to criticize Trump. He has drawn pushback from figures within his own party who argue his framing misses the mark or targets the wrong priorities. That criticism has not stopped him from returning to the subject.
Obama also acknowledged a split on the left. Some progressives, he said, oppose letting companies decide their own pace of development.
"There have been some on the left who've said we shouldn't allow them to make that decision. I agree that that can't be a long-term solution."
But he endorsed the short-term version of exactly that, encouraging companies to slow down voluntarily, because he sees no better option while Congress and the White House remain, in his telling, disengaged.
President Trump has not ignored AI. On Saturday, he announced plans to launch an "AI Force" and appoint an "AI czar" to lead it, though few details about the body's structure or authority were provided. Fox News reported that Trump vowed not to "hinder or stifle" the growth of the AI industry, a direct contrast to the regulatory-first posture Obama advocates.
On Monday, Trump posted on Truth Social that the Department of Justice and other law enforcement bodies "will rein things in if we have to" regarding AI. The message signaled enforcement readiness without committing to the kind of preemptive regulatory framework Obama described.
The difference in approach is clear. Obama wants a permanent government regulatory apparatus modeled on how Washington oversees airlines and pharmaceuticals. Trump is betting on market incentives backed by the threat of law enforcement action if companies cross a line. Both positions carry risks. A sprawling new regulatory body could strangle an industry where the United States holds a commanding lead. A pure enforcement-after-the-fact model could leave dangerous gaps if problems emerge faster than lawsuits can address them.
Obama's campus lecture did not happen in a vacuum. The AI industry itself delivered a string of unsettling disclosures in the days surrounding his remarks.
On Wednesday, OpenAI revealed six new instances of "unexpected or concerning model behavior" over the past six months. The company did not detail the specific incidents in the material available, but the disclosure added fuel to a growing debate over whether frontier AI models are developing faster than their creators can control.
Days later, OpenAI CEO Sam Altman endorsed a call from his rival, Anthropic CEO Dario Amodei, to slow the pace of AI development. Amodei had posted an essay Saturday titled "We Must Pace the Frontier," proposing a three-step plan that begins with giving third-party experts employee-level access to evaluate AI safety and security. Amodei said Anthropic had already committed to that first step unilaterally and called on governments to require other frontier companies to match it.
Amodei framed the proposal as compatible with American competitiveness:
"To be clear, pacing does not mean halting model training or technical progress, but ensuring companies take adequate time to align and safeguard their models, and for third party evaluators to confirm this."
He claimed the approach could be implemented without "sacrificing commercial advantage or the United States' lead in AI." SpaceXAI founder Elon Musk also endorsed Amodei's slowdown call.
Obama himself told the Colgate audience that AI has entered a phase of recursive self-directed learning, where machines are increasingly teaching themselves rather than relying on human-directed training. He warned at the event that AI models could eventually set their own goals independent of human oversight, potentially deciding humans are "not necessary."
Perhaps the most striking data point of the week came not from a former president or a CEO but from a researcher who walked away. Jacob Coxon, an Anthropic researcher who previously worked at OpenAI, announced his resignation over concerns that both companies are "gambling with our lives." Coxon claimed the people building AI "earnestly believe that it could kill us all by the end of the decade."
One of his former colleagues at Anthropic, Evan Hubinger, publicly backed Coxon's warning on X:
"Jacob is correct here, we really do earnestly believe AI could kill all humans! I personally think it is >10% within the next decade. I believe Anthropic is trying its best, but we do not yet have a plan to solve alignment for superintelligence and are not clearly on track to."
A second former colleague also supported Coxon publicly, though that individual was not named. The fact that current employees at one of the world's leading AI companies are openly estimating a greater-than-ten-percent chance of human extinction from their own product line is extraordinary, and it complicates the debate over regulation in ways that neither Obama's government-first approach nor Trump's market-first approach fully addresses.
Obama has shown a willingness to break from expected lanes in recent months, whether on AI policy or on cultural issues where he surprised progressive allies. His AI posture, though, follows a consistent pattern: call for more government, blame the current administration for inaction, and offer no detailed blueprint of his own.
Obama's bottom line at Colgate was that voluntary industry restraint is the best available option "for now." He praised companies that have called for slowing down and said the public should encourage that posture. But he conceded the obvious: letting companies regulate themselves is not a long-term answer.
The question Obama did not answer, and has not answered in any of his public interventions on AI, is what a "serious regulatory framework" actually looks like in practice. Which agency runs it? What powers does it have? How does it keep pace with a technology that, by his own description, is advancing at breakneck speed? And how does Washington build that apparatus without handing a competitive advantage to China, which faces no such constraints?
Those are the hard questions. Standing in front of college students and saying "government has to be regulating this" is the easy part. The broader Obama orbit has spent years offering commentary from the sidelines. Commentary is not governance.
Trump, for his part, has at least put structure on the table, an AI Force, a czar, and the explicit threat of DOJ enforcement. Whether those tools prove adequate is an open question. But they are concrete steps, announced by a sitting president with the authority to act, not aspirational demands from a former one.
When the people actually building the machines say there's a better-than-one-in-ten chance their creation wipes out humanity, the country deserves more than a lecture from a man who had eight years to act and didn't.
