AI CEOs Push Congress for DNA Synthesis Screening Laws

The heads of leading artificial intelligence companies are calling on Congress to establish mandatory screening requirements for synthetic DNA and RNA sales, warning that AI could enable bad actors to develop biological weapons.

Sam Altman of OpenAI, Demis Hassabis of Google DeepMind, Dario Amodei of Anthropic, and Mustafa Suleyman of Microsoft AI signed a public letter advocating for legislation that would require gene synthesis companies to vet both customers and orders. The letter was organized by the Institute for Progress and the Foundation for American Innovation, according to details first reported by WIRED.

The convergence of AI and synthetic biology

The letter acknowledges a stark reality: AI development may soon erode the knowledge barriers that have historically prevented malicious actors from obtaining biological weapons. While synthesizing DNA has been possible since the 1950s, the process is now automated and commercially available through dozens of companies worldwide.

The concern has grown more acute as costs have plummeted. In 2017, Canadian researchers used just $100,000 in mail-order DNA to reconstruct the extinct horsepox virus, demonstrating that similar methods could potentially be used to create smallpox. Gene synthesis has only become more affordable since then.

Combined with large language models, designing dangerous toxins and pathogens is now feasible, though some biological expertise would likely still be required to produce a functional virus. David Relman, a Stanford University microbiologist and biosecurity expert who signed the letter, notes that AI tools can quickly identify synthesis providers that don't screen orders and suggest modifications to evade detection systems.

Current safeguards and their limitations

Many gene synthesis companies already participate in voluntary screening through the International Gene Synthesis Consortium, formed in 2009. These firms use software to flag "sequences of concern" that could contribute to an organism's toxicity or disease-causing ability.

James Diggans, vice president of policy and biosecurity at Twist Bioscience, a gene synthesis company that signed the letter, argues that formal regulations are needed. "If you have technology that is capable of synthesizing DNA, then you should ensure that it's used responsibly," he said.

Federal guidelines introduced during the Biden administration require scientists and companies receiving federal funding to order from providers that screen purchases. A bipartisan Senate bill introduced this year would extend screening requirements to all gene synthesis providers operating in the United States.

However, screening tools have proven imperfect. A Microsoft study published last year showed that AI protein design tools generated potentially dangerous sequences that bypassed companies' screening software by suggesting new proteins with structures similar to known hazardous ones.

Why it matters

The intersection of AI and synthetic biology creates a dual-use technology challenge that existing voluntary frameworks may not adequately address. While bioterror attacks have been rare, they carry the potential for mass casualties and could spark a global pandemic. The letter represents a rare instance of AI company leaders proactively seeking regulation before a crisis occurs, suggesting the industry recognizes genuine risks in this domain.

Geoff Ralston, former president of Y Combinator and a partner at the Safe AI Fund, argues that AI labs developing biology models should implement their own user screening. Relman agrees that regulation of DNA synthesis alone is insufficient, noting that "the AI companies are going to have to step up" with additional controls.

The details were first reported by WIRED.