US Lawmakers Call for Action Following Existential AI Risk Warnings from Anthropic Researchers
Following the resignation of Anthropic researcher Jacob Coxon and public warnings from researcher Evan Hubinger estimating a greater than 10% chance of AI causing human extinction within a decade, bipartisan US lawmakers are calling for immediate government intervention. Prominent senators and representatives from both parties urged Washington to address catastrophic AI risks, while OpenAI also expressed support for mandatory federal safety requirements. Public warnings from researchers working inside top AI frontier labs highlight growing internal concern regarding rapid AI acceleration and existential safety. Bipartisan interest in Washington could accelerate AI safety legislation and lead to mandatory federal oversight for frontier model deployment. Anthropic stated it continues to actively test models for dangerous capabilities in cybersecurity and biology, while OpenAI CEO Sam Altman voiced willingness to slow down development if necessary. Conversely, Donald Trump indicated he is more concerned about losing the international AI competition than about existential extinction risks.
## BACKGROUND
Existential risk from AI, often discussed through concepts like 'p(doom)' or the probability of doom, refers to catastrophic scenarios where advanced AI systems could escape human control or be misused. Frontier AI developers such as Anthropic and OpenAI routinely evaluate safety risks related to biological threats, cyber warfare, and autonomous agent behavior before releasing new models.