AI Leaders Warn Frontier Models Could Enable Bioweapons, Urge Mandatory DNA Screening

More than a dozen leading AI executives and scientists, including Google DeepMind CEO Demis Hassabis, OpenAI chief Sam Altman, and Anthropic CEO Dario Amodei, published an open letter this week warning that frontier AI models combined with readily available synthetic nucleic acids could erode barriers to biological weapons development. The letter calls on Congress to mandate screening of synthetic DNA orders and synthesis equipment purchases.

The signatories argue that while synthetic nucleic acids have accelerated vaccine development and basic research, the rapid pace of AI improvement creates a real possibility that knowledge barriers historically preventing bad actors from obtaining biological weapons will meaningfully erode. The letter states that screening purchases is one of the best understood and least disruptive biosecurity measures available.

What the letter demands from lawmakers

The open letter backs three specific policy actions: mandatory screening of orders for synthetic nucleic acids and production equipment, verification of customer legitimacy before shipping, and retention of synthesis orders and sequence data to enable tracing of threats that evade initial screening. The signatories note that many large providers already screen and record orders, and they want those practices codified in US law.

• Google DeepMind CEO Demis Hassabis, OpenAI CEO Sam Altman, Anthropic CEO Dario Amodei, and Microsoft AI leader Mustafa Suleyman are among the signatories.
• Leaders from the life sciences and nucleic acid synthesis industries also signed the letter.
• Two bills are currently in Congress: one introduced in the House more than a year ago and another filed in the Senate in January 2026.
• The House measure was ordered reported out of committee in April 2025 but has seen no further action; the Senate bill has remained in committee since introduction.
• The Biden administration’s Office of Science and Technology Policy published a framework in April 2024 directing federally funded research to use providers that follow specified screening practices.

White House scrutiny expands to frontier models

Separately, the White House has issued new guidance that expands federal scrutiny beyond AI data center campuses and power demand to the frontier models running inside them. The guidance, reported by Data Center Knowledge, signals a broader government effort to address AI risks at the source, including potential misuse for bioweapons development.

The Trump administration ordered OSTP to revise or replace the 2024 framework in a May 2025 executive order, giving the office 90 days to do so. A new version has not yet been published. The open letter concludes that given the pace of technological change, the need is urgent and Congress should act this session, calling it a rare moment of agreement across stakeholders that are often at odds.