Core Summary
OpenAI CEO Sam Altman has publicly acknowledged that the world’s fear of artificial intelligence is “reasonable,” while stressing that the public should “trust” the safety efforts of AI companies. The remarks reflect both an honest acknowledgment of AI’s potential threats and OpenAI’s attempt to find a balance between regulatory pressure and commercial interest.
Event Details
Altman’s Latest Statements
According to the BBC, Altman admitted in a recent interview that public concern about AI risks is reasonable. Several tech CEOs echoed similar views, arguing that AI companies “have sufficient incentives to limit risks in technology development,” since any major incident would destroy the company’s commercial value.
Altman further noted that AI companies are actively accepting external oversight, including supporting congressional safety-assessment legislation. He stressed that combining AI industry self-regulation with a regulatory framework is the right path to addressing AI risks.
Supporting Policy Developments
OpenAI backs congressional legislation: According to Politico, OpenAI has publicly supported a bipartisan House bill on third-party AI safety assessments—the first time OpenAI has explicitly backed a specific regulatory bill.
CNBC reported: OpenAI CFO Sarah Friar said in a recent interview that OpenAI is not worried about AI development being “overly slowed” and stressed that “safe development” and “rapid development” are not in conflict.
Reuters added: OpenAI is also supporting bipartisan legislation on AI and bioweapon threats, aimed at preventing AI from being used to generate information on harmful biological agents.
Convergence and Divergence Across Stakeholders
Tech CEO Consensus: CEOs from OpenAI, Google, Anthropic, and Meta have all publicly acknowledged AI risks but oppose the extreme “pause-for-safety” stance.
Congressional Position: Some members of Congress are skeptical of AI companies’ self-regulatory pledges and advocate for mandatory legislation to strengthen oversight of frontier AI models.
Academic View: AI researchers at Stanford and MIT are calling for “moderate-intensity regulation”—tiered regulation at the application layer while keeping basic research open.
Panoramic Analysis
First, behind the “trust AI companies” rhetoric lies a battle over commercial boundaries. Altman’s “trust AI companies” stance is essentially OpenAI’s attempt to win “self-regulation priority” in regulatory legislation. If industry self-regulation is accepted by lawmakers as “sufficient safety assurance,” then mandatory third-party assessments, disclosures, and pauses on AI models will be significantly weakened. This battle for narrative is the core battlefield in AI regulatory legislation.
Second, AI regulation has shifted from “whether to regulate” to “how to regulate.” Compared with a year ago, the focus of the AI regulation debate has shifted from “whether” to “who regulates, who is regulated, and how far.” By actively supporting third-party safety assessments, OpenAI is trying to shape a “self-compliance by AI companies” paradigm that avoids the more radical “government-mandated licensing” model.
Third, AI’s “biosecurity risk” is becoming the new legislative focus. OpenAI’s active support for legislation on AI and bioweapon threats reflects that AI companies themselves recognize “biosecurity” as the regulatory entry point most likely to be accepted by lawmakers. Once such legislation lands, it will create hard compliance obligations for AI companies to assess biosecurity risks, forcing the entire industry to build stricter safety review mechanisms.
Fourth, the “trust” narrative is displacing the “fear” narrative. A year ago, the dominant AI regulation narrative was “AI could destroy humanity,” calling for a pause on frontier AI training. Today, Altman’s “trust AI companies” narrative is replacing the fear narrative. This shift reflects the success of AI companies’ PR efforts over the past year and means regulatory legislation may move toward “soft” forms.
Multiple Perspectives
OpenAI’s Position: Altman emphasizes “AI risks are real, but can be managed through responsible development.” He calls for a “third-party safety assessment system for frontier AI models” and opposes the extreme position of trading safety for a pause.
Anthropic’s Position: CEO Dario Amodei has long championed a “responsible scaling policy,” calling for government-level licensing of training and deployment of ultra-large models—a more aggressive stance than OpenAI’s.
Nvidia’s Position: Jensen Huang opposes dedicated legislation for the AI sector, advocating reliance on existing antitrust, consumer-protection, and data-privacy laws. This position is in stark contrast to OpenAI and Anthropic.
Government Position: The White House Office of Science and Technology Policy has issued the “AI Bill of Rights Blueprint,” but it lacks legislative force. The EU AI Act has entered the implementation phase and will impose real constraints on global AI companies.
Academic Position: Most scholars at institutions such as Stanford’s HAI support “tiered regulation”—strict regulation for high-risk AI applications and an open attitude toward low-risk ones.
Editor: GoodInfo Global News Team