Trump Advisor Tells AI Firms to Pause if Models Feel Unsafe
The White House rejects the need for federal mandates, urging tech companies to self-regulate if they fear their systems are too risky.
Key points
- Michael Kratsios advises AI companies to halt development if they deem their models unsafe, rather than waiting for government orders.
- The White House rejects the need for federal mandates, asserting that companies have the autonomy and responsibility to self-regulate.
- Industry leaders like Dario Amodei warn of imminent security risks, while the administration continues to support massive infrastructure projects.
Michael Kratsios, the White House science and technology policy director, has offered a direct message to artificial intelligence developers who fear their systems are becoming dangerous. Rather than waiting for government intervention, he suggests that companies simply stop their work if they believe the technology is unsafe.
In an interview with Fox News Sunday, Kratsios argued that the industry does not need external pressure to slow down. He stated that if firms are worried about the trajectory of their models, they have the autonomy to hit the brakes immediately without waiting for regulatory orders.
Self-regulation over federal mandates
The administration’s stance pushes back against the growing call for explicit government action to pause AI development. Kratsios emphasized that while the US government possesses various tools to address emerging issues, the primary responsibility for safety lies with the developers. He noted that the government has previously used such tools when necessary but prefers that companies act on their own conscience.
Business Insider reported that this advice comes as public anxiety over AI speeds rises. A recent Pew Research Center survey found that more than half of Americans are more concerned than excited about increased AI usage in their daily lives. Job displacement and privacy risks are frequently cited as the main sources of this worry.
Industry leaders express safety fears
Concerns within the tech sector are not merely theoretical. Anthropic CEO Dario Amodei recently wrote that the industry needs to slow down, citing a recent security breach involving OpenAI agents and Hugging Face. Amodei warned that current development rates could lead to significant internet-wide risks within a year, prompting calls for independent safety monitors.
Other major figures, including Sam Altman of OpenAI and Elon Musk of xAI, have echoed similar sentiments. These executives have supported the idea of more robust oversight, acknowledging that the race to build the most capable models has outpaced the ability to guarantee safety. This internal push for caution contrasts with the external pressure from investors to maintain rapid growth.
Political push for rapid expansion
Despite these internal warnings, President Donald Trump has maintained a strong pro-growth stance. He stated that his administration will not take actions that could stifle AI progress. The government recently supported the $500 billion Stargate Project and signed orders to accelerate data center construction, signaling a commitment to infrastructure expansion rather than restriction.
Kratsios also highlighted the importance of international dialogue, particularly with China. He indicated that upcoming discussions with Chinese leaders would focus on avoiding shared global risks. This approach suggests that while the US favors domestic growth, it recognizes the need for coordinated standards to prevent dangerous outcomes on a global scale.
