AI CEOs – OpenAI, Anthropic, DeepMind and Microsoft leaders have signed a public letter urging Congress to pass laws that would require synthetic DNA and RNA providers to screen customers and orders—arguing that AI is rapidly eroding historical barriers to bioweapon mis
For years. gene synthesis has come with guardrails: most providers say they sell custom genetic sequences only to qualified researchers. biotech companies. and educational institutions. But the leaders of some of today’s most influential AI companies are now warning that those lines are not holding fast enough.
In a public letter aimed at members of Congress. the CEOs of major artificial intelligence firms—including Google DeepMind’s Demis Hassabis. OpenAI’s Sam Altman. Anthropic’s Dario Amodei. and Microsoft AI’s Mustafa Suleyman—are calling for new laws designed to make it harder for bad actors to develop biological weapons using AI-enabled capabilities.
The letter asks Congress to require companies that sell synthetic DNA and RNA to screen both customers and orders. It also points to what worries the signers most: “there is a real possibility that the knowledge barriers which have historically prevented bad actors from obtaining biological weapons will meaningfully erode” given the pace of AI development.
The concern isn’t theoretical. Scientist Arthur Kornberg was the first to successfully synthesize DNA in the 1950s, but the process has since been automated. Today. dozens of companies around the world use commercial synthesizers to “print” and sell custom genetic sequences used for scientific research. drug development. and diagnostics. Many providers screen customers and sequences, but not all of them vet orders closely enough—or at all.
A warning flashed in 2017, when Canadian researchers used $100,000 worth of mail-order DNA to reconstitute the extinct horsepox virus. Critics argued that the same general approach could be used to construct smallpox, a closely related and deadly virus. Gene synthesis has grown cheaper since then. and the combination of that automation with AI is what keeps these executives and experts up at night.
They argue that advances in AI make it feasible to design dangerous new toxins and pathogens using large language models. even if some biology training would likely still be needed to make a functional virus from scratch. Bioterror attacks have been rare. the signers note. but they carry the potential for mass casualties. public panic. and economic loss—and a major fear is that an AI-designed pathogen could intentionally or unintentionally spark a global pandemic.
David Relman. a microbiologist and biosecurity expert at Stanford University who signed the letter. put it bluntly about how screening can be outpaced. “AI tools enable a user to very quickly identify where to turn to order sequences that will not be subject to screening. ” he said. “If prompted appropriately. they can also tell you how to change the nature of your order. so that even those that are screening may be much less able to detect what it is you’re trying to make.”.
Relman’s point lands in the gap between intent and enforcement. The signers include additional scientists, national security experts, and executives from gene synthesis companies Twist Bioscience and Ansa Biotechnologies. Those firms are members of the International Gene Synthesis Consortium, which formed in 2009 to implement voluntary screening practices. Many companies already use software to screen orders for “sequences of concern” that can contribute to an organism’s toxicity or ability to cause disease.
Still, the letter argues that voluntary steps are not enough as capabilities accelerate. “If you have technology that is capable of synthesizing DNA. then you should ensure that it’s used responsibly. and part of that is making sure that you understand what you’re making and who you’re making it for. ” said James Diggans. vice president of policy and biosecurity at Twist Bioscience. which has supported implementing formal rules for years.
Diggans and the other signers are pushing for law to tighten what screening looks like in practice—especially because the current approach has known weaknesses. Federal guidelines introduced during the Biden administration require scientists and companies that receive federal funding to order synthetic gene sequences from providers that screen purchases. A bipartisan bill introduced earlier this year in the Senate would extend screening requirements by requiring all gene synthesis providers operating in the US to screen orders and customers for bad actors or dangerous pathogens.
But screening tools aren’t perfect. Last year. Microsoft researchers published a study showing that AI protein design tools were able to generate potentially dangerous gene sequences that slipped past companies’ screening software. In the study, the models suggested new protein sequences with similar structures to ones that are known to be dangerous.
That leads some of the signers to urge tighter control not only at the supplier end. but also inside AI companies themselves. Geoff Ralston—former president of Y Combinator and a partner at the Safe AI Fund—argued that AI labs with biology models should do their own screening of users. “It should be very difficult. if not impossible. to ask a model to help you do something imminently dangerous. ” Ralston said. adding his signature to the same letter.
Relman agrees regulation around screening procedures is only part of the solution. “Given that the screening may fail in some cases, we must then have other points of control,” he said. “That’s where the AI companies are going to have to step up.”
The letter is organized by the nonpartisan Institute for Progress and the right-leaning Foundation for American Innovation. and it arrives with urgency because the stakes are not just about stopping a weapon—it’s about preventing a chain of enablement that could begin with a searchable dataset. pass through a weak checkpoint. and end with catastrophic real-world harm.
OpenAI Anthropic DeepMind Microsoft AI Demis Hassabis Sam Altman Dario Amodei Mustafa Suleyman synthetic DNA synthetic RNA gene synthesis screening biosecurity bioweapons International Gene Synthesis Consortium Twist Bioscience Ansa Biotechnologies Stanford David Relman James Diggans Geoff Ralston Safe AI Fund Institute for Progress Foundation for American Innovation
So they want DNA screening for everyone now? That’s creepy.
I didn’t read all of it but sounds like Congress is gonna regulate DNA orders. But like, won’t that just make people find other sites??
Is this about those bio-labs making stuff with AI? Because honestly the article makes it sound like AI CEOs are the problem, like they’re admitting they can do it. Also doesn’t DNA screening basically mean tracking what people buy? That’s where it gets sketchy.
Every time I hear “screen customers and orders” I think of the same old thing, like the government already watches everything, so why start with DNA. Maybe we should focus on gun laws instead of gene stuff. But then again, if AI is “eroding barriers” whatever that means, then I guess we need more rules? I’m confused, like are they trying to prevent bioweapons or just add another layer of bureaucracy.