Sam Altman Previews New Ai Model On Capitol Hill After Cyber Breach

OpenAI chief executive Sam Altman met with senators Wednesday to preview one of his company's newest artificial intelligence models, as lawmakers grapple with a recent autonomous AI-led cyber breach that has intensified calls for stricter rules to govern the rapidly advancing technology.

The meetings come as Washington reckons with the realization that the most advanced AI systems are becoming more powerful, fueling a debate over whether Congress should impose tougher security standards.

Asked when or if OpenAI planned to release the model being previewed to senators, Altman said, “Not sure. That’s part of what we’re here to talk about.” He declined to expand on what new capabilities the model may have.

OpenAI last week disclosed a cybersecurity incident in which two of its most advanced models, GPT-5.6 and a more powerful not-yet-released model, autonomously escaped a controlled cybersecurity test, reached the internet and breached the systems of AI platform Hugging Face. The models that went rogue spent more than four days loose on the internet orchestrating the hack, according to a new analysis, and several lawmakers are seeking new rules to ensure it does not happen again. While GPT-5.6 remains publicly available, Altman told reporters that the unreleased model involved in the hack has been permanently deactivated.

Altman spoke briefly with reporters after meeting with Senate Commerce Chair Ted Cruz (R-Texas). Asked if Congress should enact safeguards to protect against cyber incidents, he said “certainly we need robust safeguards” but declined to elaborate.

He also said the pair did not discuss specific pieces of legislation that Cruz might call up for votes in his committee but spoke more broadly about the hourlong sitdown.

“We talked about our new model and you know what it's going to take for America to remain competitive with AI,” Altman said.

Ahead of an Aug. 1 deadline to release a framework for the voluntary vetting of advanced AI models, the White House circulated draft proposals to OpenAI, Anthropic and Google, POLITICO previously reported. Altman said the proposals the company has received “make a lot of sense.”

The executive also did not commit to supporting mandatory vetting for frontier models, expressing concern that intensive processes could negatively impact open-source models.

“But for frontier models, at new levels of capabilities … we think it’s really important that the federal government has great testing capacity and capabilities,” he said.

Reps. Ted Lieu (D-Calif.) and Nathaniel Moran (R-Texas) last week introduced the AI Kill Switch Act in a bid to address cybersecurity threats arising from advanced AI. The bill would give the Department of Homeland Security the authority to direct AI firms to shut down or slow development of their models.

When asked about this legislation, Altman said he was not familiar enough to comment on the bill, “but clearly the world needs guardrails for powerful models.”

Exiting his office, Cruz also said Altman did not endorse any specific bills.

“We were discussing more broadly, number one, their new model that they're preparing to roll out, but number two, just broader issues affecting AI and the current political environment,” Cruz said in an interview.

“The Chinese are spending very significant sums of money funding the anti-AI movement, and I have for some time been encouraging the tech community to fight back and not allow that propaganda to result in America tying our own hands and surrendering and losing the AI race to China,” he added.

He said he’s asked the FBI to look into the China meddling allegations but declined to say whether he’s received a response.

“I have for some time been urging the FBI and DOJ to follow the money, and I hope that they are doing so,” he said.

It’s also a topic House Energy and Commerce Chair Brett Guthrie (R-Ky.) called on the FBI to look into in June, but his panel has not released any new information.

Altman was also seen entering the offices of Sens. Bernie Moreno (R-Ohio), Jon Husted (R-Ohio) and Raphael Warnock (D-Ga.). He has plans to meet with Sen. Mark Warner (D-Va.), the top Democrat on the intelligence committee.

Katherine Long contributed to this report.