AI CEOs Urge Congress to Mandate Biosecurity Screening for Gene Synthesis
Summary
- • Altman, Amodei, Hassabis, and Suleyman sign joint letter urging bioweapons screening laws
- • Letter calls for mandatory customer and order screening by all synthetic DNA/RNA providers
- • AI tools can now help bad actors identify unscreened gene sequence suppliers
- • Microsoft research showed AI protein design tools generate sequences that evade screening software
Details
Top AI CEOs sign public letter urging mandatory biosecurity laws for gene synthesis providers
Sam Altman (OpenAI), Dario Amodei (Anthropic), Demis Hassabis (Google DeepMind), and Mustafa Suleyman (Microsoft AI) are among the signatories. The letter was organized by the nonpartisan Institute for Progress and the right-leaning Foundation for American Innovation, and also includes scientists, national security experts, and executives from gene synthesis companies Twist Bioscience and Ansa Biotechnologies.
Letter calls for laws requiring all synthetic DNA/RNA sellers to screen customers and orders
Current federal guidelines introduced under the Biden administration only require screening by providers serving federally funded researchers. A bipartisan Senate bill introduced earlier this year would extend mandatory screening to all gene synthesis providers operating in the US, covering both customers and the gene sequences they order.
AI tools can help bad actors locate unscreened suppliers and obscure dangerous sequence orders
Stanford microbiologist and biosecurity expert David Relman, a letter signatory, notes that AI tools can rapidly identify suppliers that do not screen orders, and can advise users on how to modify orders to evade detection even by providers that do screen. The letter warns that AI development is eroding the knowledge barriers that historically prevented bioweapons development.
Gene synthesis is cheap and widely available, with a history of misuse risk dating to 2017
In 2017, Canadian researchers reconstituted the extinct horsepox virus using $100,000 of mail-order DNA, raising fears the same method could produce smallpox. Since then, synthesis costs have continued to fall. Many providers already use software to screen for sequences of concern, and the International Gene Synthesis Consortium has implemented voluntary screening practices since 2009 — but not all providers vet customers or orders.
Microsoft study found AI protein design tools can generate dangerous sequences that bypass screening software
Published last year, Microsoft research demonstrated that AI models could suggest novel protein sequences structurally similar to known dangerous ones, and that these sequences slipped past commercial screening tools. This finding underscores the gap between existing screening technology and the pace of AI-assisted biological design.
Industry Update = notable sector development, Policy = regulatory or legislative action, Security Alert = identified threat or vulnerability, Context = background framing, Research = published scientific or technical findings
What This Means
The convergence of cheap gene synthesis and capable AI models has created a credible new vector for bioweapons development — one that existing voluntary screening frameworks were not designed to handle. By signing this letter, the leaders of the most powerful AI labs are publicly acknowledging that their own technology contributes to this risk and are calling for legislative guardrails rather than waiting for voluntary industry action to suffice. For policymakers, this represents a rare moment of alignment between AI industry leaders, biosecurity experts, and national security voices, potentially accelerating the bipartisan Senate bill already in play. The practical challenge, highlighted by Microsoft's own research, is that screening tools are already struggling to keep pace with AI-assisted sequence design — meaning any mandated screening regime will need to evolve continuously alongside the models it is meant to constrain.