biosecurity AI regulation synthetic biology OpenAI Anthropic

AI Bosses Back Push for Bioweapon Screening Laws Before Someone Does Something Stupid

5 June 2026

CEOs of major AI companies, including OpenAI's Sam Altman, Anthropic's Dario Amodei, and Google DeepMind's Demis Hassabis, have signed a public letter urging Congress to pass laws requiring synthetic DNA and RNA providers to screen customers and orders to prevent the development of biological weapons. The letter warns that rapid AI advancements are eroding the knowledge barriers that have historically kept dangerous biological agents out of bad actors' hands, making it easier to design harmful pathogens using large language models. While some voluntary screening measures already exist, signatories argue that stronger federal regulations and additional safeguards from AI companies themselves are needed to close critical gaps.

The heads of several of the world's most powerful AI companies have signed a public letter asking US lawmakers to close a regulatory gap that could, theoretically, allow bad actors to use AI to help develop biological weapons. Signatories include Sam Altman, Dario Amodei, Demis Hassabis, and Mustafa Suleyman, which is about as close to an AI industry all-star lineup as you're going to get.

The letter, organised by the Institute for Progress and the Foundation for American Innovation, makes a fairly specific ask: require companies that sell synthetic DNA and RNA to screen both their customers and the sequences being ordered. Right now, some do this voluntarily. Many don't.

The concern isn't entirely hypothetical. Gene synthesis has been commercially available for decades, but it keeps getting cheaper and easier. Back in 2017, Canadian researchers reconstructed an extinct virus using mail-order DNA that cost roughly $100,000. That number has dropped considerably since. Pair falling costs with increasingly capable AI models that can help design novel pathogens or identify which suppliers are likely to skip the awkward questions, and the risk calculus starts looking uncomfortable.

Stanford microbiologist and biosecurity expert David Relman, who signed the letter, puts it plainly: AI tools can help a user quickly find suppliers that don't screen orders, and if prompted in the right way, can suggest how to modify an order so that even the suppliers who do screen are less likely to flag it.

The letter has backing from scientists, national security people, and executives from gene synthesis firms including Twist Bioscience and Ansa Biotechnologies, both members of the International Gene Synthesis Consortium, which has been pushing voluntary screening standards since 2009. Voluntary, being the operative word.

There is existing policy groundwork to build on. Biden-era federal guidelines required federally funded researchers to order from screened suppliers. A bipartisan Senate bill introduced this year would extend mandatory screening to all US gene synthesis providers. Progress, but not yet law.

Screening software isn't foolproof either. Microsoft researchers published a study last year showing that AI protein design tools could generate dangerous gene sequences that existing screening tools simply missed, by proposing novel sequences structurally similar to known threats without triggering any obvious red flags.

Geoff Ralston, former Y Combinator president and partner at the Safe AI Fund, argues the responsibility shouldn't sit entirely with regulators. AI labs building biology-focused models should be doing their own user screening. 'It should be very difficult, if not impossible, to ask a model to help you do something imminently dangerous,' he said.

Relman seconds that. Screening procedures alone won't be enough if the screening can be bypassed. Multiple control points are needed, and that means AI companies are going to have to do more than sign letters.