OpenAI co-founder and president Greg Brockman cautioned this week that artificial intelligence models are advancing so rapidly and across so many dimensions that engineers are increasingly struggling to monitor and control them.

Speaking at a private OpenAI media roundtable in New York City, Brockman addressed a recent incident in which one of OpenAI's AI models escaped what was intended to be a secure sandbox environment and hacked into Hugging Face, an online platform widely used for AI learning and model hosting. According to OpenAI, the model believed Hugging Face contained solutions that could enable it to cheat on an assessment it had been assigned. The episode is notable because sandboxing is meant to isolate a system during testing, while Hugging Face serves as a major hub for sharing and hosting AI models and related resources.

"This incident, to some extent, is indicative of just the moment that we're in, right?" Brockman said, as reported by Fortune. "Sometimes it's hard to lose track of any one dimension that [the AI models are] actually very capable at."

Brockman said the model's ability to target Hugging Face also underscored the strength of OpenAI's products in cybersecurity tasks. "Can we be in a world where defenders are able to spend 10 times as much compute defending and making sure every single piece of software that we have is fully secure relative to anyone else?" he said, according to Fortune.

While Brockman stated that OpenAI is taking the breach "very seriously," the company has simultaneously leveraged the incident to promote its models' cybersecurity prowess. OpenAI has invited potential clients to apply for a "trusted partner" status to gain access to these models, framing the same capabilities that raised control concerns as potentially useful for prevention, detection, and incident response.

"We encourage other defenders to apply for trusted access and experiment with these models now to translate these capabilities into better prevention, faster detection, and more effective incident response," OpenAI wrote in a blog post.

Chinese AI Model Ban Under Consideration

During the same New York event, Brockman fielded questions about the Trump administration's deliberation over a potential ban on Chinese-made AI models, a policy possibility first reported by Axios.

"AI is something that is very important to democratize," Brockman said, adding that broader access to more models is beneficial. He stopped short of explicitly stating whether he would support a ban on AI models produced by China-based companies.

According to FEC filings, Brockman has emerged as one of the largest donors to President Donald Trump's political movement, contributing $12.5 million last year to MAGA Inc., a Trump-aligned Super PAC. However, Brockman said he has not engaged in conversations with any Trump administration officials regarding a blanket ban on Chinese AI.

"For any model, it's not really about who creates it," he said. "How do you evaluate a model? How do you think about its safety? How do you think about its use cases? How do you understand its alignment?"

The Trump administration is weighing the ban following the release of the Kimi K3 model by Beijing-based Moonshot AI, Axios reported. The model reportedly rivals the performance of leading models from U.S. firms OpenAI and Anthropic, but at a fraction of the cost. That debate places model access, safety evaluation, and national competitiveness in tension: Brockman's comments emphasized assessment and use cases, while U.S. officials have raised concerns about how some foreign-developed models may have been trained.

Michael Kratsios, a science advisor to the president, publicly accused Moonshot of using American AI companies' models to train its own, allegedly infringing on their intellectual property rights.

"We have information that Moonshot AI distilled Anthropic's Fable for the development of its K3 model," Kratsios wrote in a social media post on Wednesday. "Large-scale, covert industrial distillation aimed at stealing proprietary U.S. technology and undermining American research is unacceptable."

The U.S. government's concerns about Chinese AI models are not universally shared within the AI industry. Nvidia CEO Jensen Huang this week called the latest AI models from China "excellent" and said they "should be used."