Sam Altman Hearing: 2023 vs 2025 Senate Testimony on AI

Across two appearances two years apart, Sam Altman’s Senate testimony on AI regulation moved from asking Congress to license powerful AI models to warning that pre-release government approval would be “disastrous.” In May 2023, the OpenAI chief executive told a Senate Judiciary subcommittee that a new federal agency should license and, if necessary, revoke permission to operate the most capable AI systems. In May 2025, he told a Senate Commerce Committee focused on beating China that the country needed “sensible regulation that does not slow us down.”

What Altman Asked Congress to Do in 2023

Altman appeared on May 16, 2023, before the Senate Judiciary Subcommittee on Privacy, Technology, and the Law, chaired by Senator Richard Blumenthal. The hearing was titled “Oversight of A.I.: Rules for Artificial Intelligence.”1U.S. Senate Judiciary Committee. Oversight of A.I.: Rules for Artificial Intelligence Altman told senators that government intervention “will be critical to mitigate the risks of increasingly powerful” AI systems.2PBS NewsHour. OpenAI CEO Sam Altman Testifies Before Senate Judiciary Committee

His concrete proposals included:

Altman defended releasing models publicly as a way to gather real-world feedback, and he acknowledged that AI would displace jobs while raising productivity, saying OpenAI was funding research into policy responses such as modernized unemployment insurance.3U.S. Senate Judiciary Committee. Testimony of Samuel Altman Asked by Senator Lindsey Graham whether Section 230 covered OpenAI, Altman said he did not believe Section 230 was the “right framework” for generative AI.4Brookings Institution. Senate Hearing Highlights AI Harms and Need for Tougher Regulation

One caveat surfaced during the same hearing. Gary Marcus, a co-panelist and professor emeritus at NYU, pointed out that while OpenAI publicly supported independent audits and advance reviews, the company had not actually submitted to them.5U.S. Senate Judiciary Committee. Testimony of Gary Marcus

What Changed Between the Two Hearings

The 2023 hearing kicked off a busy stretch in Washington. In the fall of 2023, Senate Majority Leader Chuck Schumer convened nine closed-door “AI Insight Forums” with more than 150 experts from industry, academia, labor, and civil society.6Tech Policy Press. US Senate AI Insight Forum Tracker On October 30, 2023, President Biden signed Executive Order 14110, which directed NIST to establish AI safety guidelines and red-teaming benchmarks and used the Defense Production Act to require developers of large foundation models to report training activities, cybersecurity protections, and red-teaming results to the federal government.7Federal Register. Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence

Congress did not pass comprehensive AI legislation during the 118th Congress. The Insight Forums produced a policy roadmap calling for increased federal R&D funding, pre-deployment evaluations of high-risk models, content watermarking, and data privacy legislation, but none had been enacted into law by the time Altman returned to the Senate.8Office of Senator Chuck Schumer. Bipartisan Senate AI Policy Roadmap

What Altman Told the Senate in 2025

Altman’s second appearance came on May 8, 2025, before the Senate Commerce, Science, and Transportation Committee, chaired by Senator Ted Cruz. The hearing was titled “Winning the AI Race: Strengthening U.S. Capabilities in Computing and Innovation.”9U.S. Senate Commerce Committee. Winning the AI Race: Strengthening U.S. Capabilities in Computing and Innovation The framing had shifted from AI risks to competition with China, and Altman’s regulatory posture moved with it.

Altman told senators that requiring developers to obtain government approval before releasing AI systems would be “disastrous” for the country’s technology lead.10Washington Post. Altman Congress OpenAI Regulation Asked about proposals for NIST to set AI standards, he responded, “I don’t think we need it. It can be helpful.”11Tech Policy Press. Transcript: Sam Altman Testifies at US Senate Hearing on AI Competitiveness He called for “sensible regulation that does not slow us down” and aligned himself with Cruz’s “light touch” approach modeled on the hands-off internet policies of the 1990s.

Most of Altman’s 2025 testimony focused on infrastructure rather than safety rules. He told senators that “infrastructure is destiny” and pointed to the Stargate Project, a $500 billion partnership among OpenAI, SoftBank, Oracle, and others to build AI data centers across the United States, anchored by a flagship facility in Abilene, Texas.12Congress.gov. Winning the AI Race Hearing Transcript13OpenAI. Announcing the Stargate Project He praised Texas for being “unbelievable” in its efforts to attract AI investment while cautioning against a “patchwork regulatory framework” of 50 different state-level AI laws.14PBS NewsHour. OpenAI Co-Founder Sam Altman Testifies on AI Competition in Senate Hearing

Cruz used the hearing to announce forthcoming legislation creating an “AI regulatory sandbox.” That bill, the SANDBOX Act (S.2750), was introduced on September 10, 2025, and referred to the Commerce Committee. It reflected the hearing’s deregulatory tone, providing legal clarity for AI developers without the pre-release approvals Altman had warned against.15Congress.gov. S.2750 – SANDBOX Act

How the Washington Post and Others Read the Reversal

The Washington Post characterized Altman’s 2025 testimony as part of a broader industry shift, noting that “AI execs used to beg for regulation” and now appeared to be running from it.10Washington Post. Altman Congress OpenAI Regulation Some observers suggested the initial support for regulation may have functioned as a competitive strategy to raise barriers for smaller entrants, and that leading companies shifted to prioritize growth as commercial stakes rose.

Where OpenAI’s Position Landed Afterward

By mid-2026, OpenAI had opened a policy workshop in Washington and published a framework it called “Industrial Policy for the Intelligence Age,” arguing that “incremental policy updates won’t be enough” as the industry moves toward superintelligence.16OpenAI. Industrial Policy for the Intelligence Age

On oversight, the company settled into a middle position. OpenAI executive Chris Lehane said the company continued to support mandatory evaluations of advanced AI models, stating that “we don’t think any specific lab should be making that decision unilaterally.” But OpenAI argued those evaluations should sit with the Center for AI Standards and Innovation, a civilian office inside NIST, rather than with the National Security Agency, which a 2026 Trump executive order had tasked with evaluating AI cybersecurity risks. The company opposed any requirement for a government “green light” before model releases, calling that step “too early.”17Politico. OpenAI White House AI Safety Rules

That is a narrower version of what Altman first asked for. In 2023 he wanted a licensing agency with the power to shut noncompliant companies down; by 2026 OpenAI supported mandatory evaluations but rejected pre-release approval. Whether the shift reflects a change in Altman’s thinking about what regulation would accomplish, a strategic response to the competitive landscape, or the difference between asking for rules in the abstract and facing them as a market leader is a question his two Senate appearances raise but do not answer.