A Rogue Hack and China at the Door Spark a Great AI Panic
Pressure builds in Washington to rein in AI, while the industry fights over who should hold the leash
AI panic has swept Washington.
And, in real time, we’re watching wannabe masters of the artificial-intelligence universe try to win an upper hand in controlling the new technology.
OpenAI’s revelation this past week that its AI models escaped confinement and hacked another tech company’s systems gave new credibility to fears that this technology can, in fact, go sci-fi villain mode.
Add into the mix announcements that Chinese AI rivals have new, free models, allegedly built off the learnings of the expensive ones developed by the likes of Anthropic. That opened a new front in the geopolitical fight between China and the U.S.
The Trump administration was already staring down a deadline to put in place the mechanisms for a June 2 executive order to increase oversight of the biggest AI labs’ models. That order was sparked by fears the systems are becoming more powerful.
Taken all together in such rapid succession, it has increased the pressure in Washington to do something. The fear of many in the industry is that a rush to do something—anything—will hinder innovation or benefit one player over another.
For years, the largest AI companies have been trying to get ahead of things, advocating proactively for regulations to place guardrails around the powerful technology. They argued this was needed to prevent disaster for cybersecurity, biological threats and other risky areas.
Such efforts drew eye rolls among those who didn’t believe the predictions. Then, as the technology advanced, their calls drew criticism that what they were really trying to do was write the rules for their own benefit—regulatory capture—at the expense of smaller rivals.
Still, just days ago, it seemed like an unlikely unified front of AI leaders might finally make headway in Washington, D.C.
On July 14, Google DeepMind co-founder Demis Hassabis published a proposal calling for the creation of an AI standards body. This would be modeled after the self-regulating organization that oversees stockbrokers. Funding would come from the industry and a board that would include technical specialists and open-source representatives.
“When there is a large degree of uncertainty and the stakes are this high, proceeding with cautious optimism is the sensible and correct strategy,” Hassabis wrote. “That calls for public policy that promotes innovation while also incentivising responsibility and security, fosters international collaboration on key safety issues, and encourages careful consideration of how AI is deployed for the benefit of society.”
The Nobel laureate is a deeply respected, superstar in the AI field but has less of a public profile in the U.S., where his peers have become more polarizing figures.
Before publishing, Hassabis had worked to build support for the idea among his rivals. The effort seemed to pay off—even if it created some odd bedfellows.
Elon Musk called Hassabis’s proposal “a thoughtful framework overall and certainly a good starting point for discussion,” while OpenAI CEO Sam Altman posted something similar on X: “this is a thoughtful proposal from demis.”
You don’t see these two guys agree on much these days. Altman and Musk formed OpenAI years ago in response to Google’s acquisition of DeepMind, and both men have subsequently become bitter rivals.
Nevertheless, the enemy of your enemy’s regulation is your friend. As Musk explained in an interview with the Economist published this past week: “It is quite difficult for someone in the government who doesn’t have a deep technical understanding and isn’t driving the frontier of AI to know whether something should be released or not. However…all of the competitors have an incentive to keep the others honest.”
After publication of Hassabis’s proposal, a report from Bloomberg News suggested Treasury Secretary Scott Bessent had been developing a similar proposal. But as quickly as it seemed like they might have a breakthrough, things started to grow more complicated.
OpenAI rocked the Capitol late Tuesday with the disclosure that its models had hacked Hugging Face, an open-source AI tool company. It drew a swift reaction.
“AI is developing extremely fast with no real regulations to keep us safe,” Rep. Greg Casar (D., Texas) said in reference to the OpenAI hack. “That has to change.”
One idea to draw attention was a bipartisan bill introduced Thursday in the House that would require AI makers to include a capability within their systems to allow the government to shut down their models using a so-called kill switch.
“Humans should be in control, not machines. And when AI goes rogue, humans should have the ability to turn it off,” Rep. Ted Lieu (D., Calif.), one of the bill’s sponsors, said in a post on social media Thursday.
OpenAI’s announcement seemed a little too on the nose for some in the tech industry already suspicious of its efforts to shape regulations to its liking.
Timnit Gebru, a prominent AI researcher, on LinkedIn called the lab’s communication “top notch propaganda.”
Legendary venture capitalist Bill Gurley was especially blunt in his own social-media post: “If OpenAI violated Hugging Face in a way that ‘demands’ new regulation; let’s start with a formal criminal investigation. With a proper third party investigation. And sincere liability.”
Amid those exchanges, the Trump administration could be seen publicly wrestling with how to handle the threat of Chinese models.
A statement from Bessent suggesting Chinese AI could face trade restrictions drew pushback from some of the biggest names in tech, including the claim that he is carrying Anthropic’s water. By Friday, several tech giants issued a letter—including X posts from Microsoft CEO Satya Nadella and Nvidia CEO Jensen Huang—advocating for open AI models—a position that helps their businesses.
In fact, Huang made his first-ever X post on the matter.
Former White House AI czar David Sacks called for “panic” to stop around open-weight models, which allow anyone to download for use freely. “President Trump’s light-touch regulatory approach is working,” he posted on X. “We should remain confident in American innovation. As long as we don’t sabotage ourselves with unnecessary rules, the U.S. will continue to win.”
The irony is that as all of these tech companies wrestle to keep ahead of the quick-moving technology, they want the same thing that traditional companies desire: predictability.
Except in the AI age, no one has mastered that.