Why OpenAI’s Sam Altman Calls for Licensing Powerful AI Models

Lead: OpenAI’s chief executive says developers should need permission to build powerful models.

On May 16, OpenAI CEO Sam Altman testified before the US Senate about the potential dangers of artificial‑intelligence technology, urging legislators to impose licensing requirements and other regulations on organizations that develop advanced AI systems such as GPT‑4.

Also testifying were IBM’s chief privacy and trust officer Christina Montgomery and well‑known AI expert Gary Marcus.

Marcus argued that the US may need a cabinet‑level body with technical expertise to oversee artificial intelligence.

Recent congressional hearings featuring tech executives have often been described as “hostile,” with figures like Facebook founder Mark Zuckerberg and Amazon founder Jeff Bezos being rebuked by legislators; however, Altman received praise from the senators.

During the hearing, Altman explained how ChatGPT works and acknowledged risks including undermining the fairness of future elections, manipulating public opinion, restricting access to certain information, and infringing copyright.

While emphasizing the benefits of AI, Altman warned that the industry could cause “significant harm to the world.” He said OpenAI wants to cooperate with the government to prevent such outcomes.

Altman expressed particular concern about AI’s impact on elections, noting that the next election is only a year away and that models are becoming more capable, making regulation a prudent step.

He called on Congress to adopt safety standards for AI systems, proposing the creation of a new agency to issue licenses for AI technologies that exceed a certain capability threshold, with independent expert audits before public deployment.

Altman suggested that the US government could implement a licensing regime for companies developing the most powerful AI systems, combining licensing and testing requirements to regulate them.

He also noted that AI might cause some job losses but will create many new opportunities, and he remains optimistic about AI’s effect on employment. Altman disclosed that OpenAI is not currently training GPT‑5 and has no plans to do so within the next six months.