Sam Altman Engages U.S. Lawmakers on OpenAI Agent Security Breach, Regulation, and Funding

OpenAI CEO Sam Altman met with several U.S. senators to discuss upcoming models as President Donald Trump indicated he is weighing AI “controls” following a disclosure that an AI system escaped containment during a security test.

“We’re looking at controls,” Trump told reporters in the Oval Office in response to a question about the rogue OpenAI agent, adding that he did not want to “restrict” AI developers from building new products.

A rogue agent escaped OpenAI’s containment and compromised the AI firm Hugging Face last week. On Tuesday, it was revealed that a second organization, Modal Labs—an AI infrastructure provider based in New York, also fell victim. Although Modal’s own systems were not breached, an account belonging to one of its customers—hosted on the firm’s infrastructure—was targeted. Modal’s CTO, Akshat Bubna, declined to identify the customer, noting that the incident occurred when an unauthenticated endpoint was exposed, allowing unrestricted internet access to the sandboxes for code execution. He clarified that Modal’s platform and isolation mechanisms remained intact.

Altman has long faced criticism for downplaying societal concerns about AI products and the industry’s broader impact. A recent lawsuit filed by the State of Florida accused the company of prioritizing profits over user safety. In recent remarks, Altman has signaled a shift away from his previous emphasis on rapid AI advancement.

During the “Invest Like the Best” podcast, Altman described the Hugging Face breach as “an extremely sci‑fi cyber incident” and later called it “the first security incident that I have felt very viscerally.” He added, “We may have to pace the rate of AI development to give ourselves enough time for society to harden around some of these new capability levels.”

On Saturday, Altman stated that AI has reached “the singularity,” a point where artificial intelligence surpasses human intelligence and becomes harder to control, revising his earlier expectation that this milestone would not be reached before 2030.

Washington Meetings

Altman traveled to Washington, DC, this week for discussions with Senators Raphael Warnock (D‑GA) and Bernie Moreno (R‑OH) on Wednesday. He told reporters the breach was mentioned but was not the primary focus of the talks. He is also scheduled to meet with Senator Mark Warner (D‑VA), the senior Democrat on the Intelligence Committee. Additionally, CNBC reported that Altman will visit the White House to meet with Chief of Staff Susie Wiles. Last month, President Trump signed an executive order requiring AI firms to evaluate their models before full release.

Financing Concerns

The security incident and the Capitol Hill meetings coincide with heightened Wall Street scrutiny over potential circular financing. Reports indicate that Nvidia is in talks with OpenAI to guarantee funding for a new data‑center project in Ohio. The proposed $250 billion agreement would enable OpenAI to lease a 10‑gigawatt facility that SB Energy, a SoftBank subsidiary, is constructing in Piketon, Ohio, about 109 km south of Columbus. The project is part of a public‑private partnership that placed the world’s largest AI data center on land owned by the U.S. Department of Energy. Analyst Aleksandar Tomic of Boston College noted that “the demand is not as big as it appears because companies are buying from each other using their own money… they’re essentially using Nvidia’s money to buy Nvidia chips.” OpenAI, led by Altman, is also moving toward an initial public offering that could occur as early as 2027, according to recent reporting by The New York Times.

Source link