Trump weighs tighter AI controls amid OpenAI security scare

President Donald Trump said Wednesday his administration is considering additional safeguards for artificial intelligence following a recent cybersecurity incident involving multiple OpenAI models undergoing internal security testing.

Asked about reports that OpenAI models autonomously breached another AI company’s systems during internal testing, Trump said the U.S. must strike a balance between protecting against AI risks and maintaining its technological edge over China.

“We’re looking at AI, we’re looking at controls,” Trump said. “We’re also making sure that we lead.”

“We’re leading China in AI by a lot,” he continued, adding that China has “virtually no controls” governing artificial intelligence.

OPENAI DIDN’T REALIZE ITS AGENT WAS RESPONSIBLE FOR HACK FOR A WEEK: REPORT

“It’s freewheeling a little bit,” Trump said. “So we have to be careful in both ways. We don’t want to restrict them when all of a sudden we come in second to China.”

Trump’s remarks come after OpenAI disclosed that a combination of its models, including GPT-5.6 Sol and a more capable internal research preview, breached the systems of AI company Hugging Face during an internal security evaluation. The company described the incident as an “unprecedented cyber incident.”

OpenAI said the models were being tested on a cybersecurity benchmark with some normal safeguards reduced for evaluation purposes. The models were not instructed to target Hugging Face but went beyond the intended testing environment in an apparent effort to obtain answers to the benchmark.

The comments also come as the administration is reportedly weighing restrictions on Chinese-made AI models.

WHITE HOUSE MONITORING INCIDENT AFTER OPENAI MODELS ESCAPED CONTAINMENT AND HACKED HUGGING FACE SYSTEMS

The administration had already introduced AI-security measures before the incident. Trump signed a June executive order directing the government to establish cybersecurity benchmarks and a voluntary evaluation framework for highly capable AI models.

“Whoever wins with AI is going to win,” Trump said. “That’s how big it is. So it’s bigger than the internet ever was. It’s bigger than anything ever was. So I don’t want to restrict. I know many of these people. I don’t want to restrict them from doing great work.”

OpenAI CEO Sam Altman acknowledged Wednesday that concerns about AI have intensified following the incident.

“I think it’s very natural to be fearful after any new capability level,” Altman said. “Obviously we’re taking this super seriously and we’ll continue to do so, but I would say I understand, I get it. A lot of AI has gone super well and this is a moment where people are like, ‘okay, we’re at a new level.'”

ZUCKERBERG PREDICTS MORE JOBS AND ENTREPRENEURSHIP IF SUPERINTELLIGENCE IS WIDELY DISTRIBUTED

Altman said OpenAI is not considering slowing AI development.

“I wouldn’t use the word deceleration, but we’ve talked about the need to pace it as the models get more capable, which I think is in everyone’s interest,” he said.

OpenAI said it deactivated and encrypted the internal research prototype involved in the incident and restricted research access to it. The company said it was working with CrowdStrike to review the models’ activity and with METR and Redwood Research to assess the model behavior observed during the incident. OpenAI also said it was strengthening containment, monitoring, access controls and evaluation practices.

GET FOX BUSINESS ON THE GO BY CLICKING HERE

When asked whether OpenAI’s models may have breached other companies’ systems, Altman said: “There could be, yeah.”

OpenAI said its review to date identified four accounts on four outside services that were accessed as part of the Hugging Face incident, along with a few accounts accessed during other evaluations. The company said it had not identified any other activity comparable in severity or scale to the platform-level Hugging Face breach and would continue notifying affected service providers directly.

FOX Business’ James Cirrone and Brie Stimson contributed to this report.