AI CEOs Urge Congress to Mandate DNA Synthesis Screening for Bioweapons Risk

AI CEOs Urge Congress to Mandate DNA Synthesis Screening for Bioweapons Risk

The chief executives of OpenAI, Anthropic, Google DeepMind, and Microsoft AI rarely agree on anything in public. On the night of June 3, 2026, they signed the same open letter to Congress, warning that artificial intelligence is eroding the knowledge barrier that has long kept biological weapons out of reach of most bad actors, and asking lawmakers to make screening of synthetic DNA orders mandatory.

The letter, titled "In Support of Mandatory Nucleic Acid Synthesis Screening and Recordkeeping," was published late on June 3 and circulated widely the following day. Among the signatories are Sam Altman of OpenAI, Dario Amodei of Anthropic, Demis Hassabis of Google DeepMind, and Mustafa Suleyman of Microsoft AI, along with Meta chief AI officer Alexandr Wang, Stripe CEO Patrick Collison, Nobel laureate and protein-design pioneer David Baker, cryptographer Martin Hellman, and executives from gene-synthesis firms including Twist Bioscience and Ansa. It was organized by two Washington policy groups, the Institute for Progress and the Foundation for American Innovation.

What the letter asks for

The core request is narrow and, the signatories argue, long overdue. They want providers of synthesized DNA and the manufacturers of synthesis machines to check every order against databases of dangerous sequences, verify that the customer is legitimate, and keep records so that biosecurity investigators can trace a suspicious order back to its source, even when the individual fragments would not raise alarms in isolation.

"As life sciences researchers, builders of AI and biotechnology, and experts with a wide range of views on how to approach AI policy, we call on legislators to make screening of orders for synthetic nucleic acids—and the equipment needed to make them—mandatory," the letter reads.

The signatories are careful to frame the case as independent of any particular view of AI risk. The biosecurity argument, they note, has been recognized by scientists and governments for decades, ever since protocols to reconstruct viruses from strands of DNA were published more than twenty years ago. The industry's own International Gene Synthesis Consortium has run voluntary safeguards since 2009. What has changed, they write, is the pace.

"AI systems now outperform PhD-level virologists on questions about highly technical laboratory procedures in their own domains of expertise," the letter states, citing a benchmark called VirologyTest. "The evidence about what this means for present-day biosecurity threats is genuinely mixed, but the trend is hard to dispute. AI systems are improving rapidly, and alongside incredible benefits to science and medicine, there is a real possibility that the knowledge barriers which have historically prevented bad actors from obtaining biological weapons will meaningfully erode."

The closing line frames the moment as unusual: "This is a rare moment of agreement across stakeholders who are often at odds. We hope policymakers will meet it with decisive action."

The bill behind it

The letter explicitly applauds "the legislative efforts currently underway," and the most relevant vehicle is S. 3741, the Biosecurity Modernization and Innovation Act of 2026. Introduced February 4, 2026, by Senators Tom Cotton (R-Arkansas) and Amy Klobuchar (D-Minnesota), the bipartisan bill would direct the Secretary of Commerce to require gene-synthesis providers to screen their orders and customers, establish a biotechnology governance sandbox at the National Institute of Standards and Technology, and order a review of fragmented federal biosecurity authorities.

"American innovations in biotechnology are too important to fall into the hands of bad actors or be hamstrung by outdated federal policies," Cotton said when introducing the measure. "This bill is an important first step towards a comprehensive framework for biosecurity and biosafety."

Klobuchar framed the dual-use problem plainly. "While access to genetic material allows scientists to study diseases, develop lifesaving medicine, and improve crops, without safety standards it could be misused, including to create bioweapons," she said. "Our bipartisan legislation will enhance biosecurity efforts and reduce the risk of genetic material being used to cause harm by ensuring that the companies selling these products know who their customers are and how they will use genetic material."

The bill has been endorsed by the National Security Commission on Emerging Biotechnology, the Engineering Biology Research Consortium, and gene-synthesis firms including Twist Bioscience, Aclid, Integrated DNA Technologies, and Ginkgo Bioworks. A parallel House measure, H.R. 3029, takes a softer, voluntary-compliance approach, setting up a likely point of friction as the two chambers try to reconcile.

Why the rivals align

It is worth sitting with how unusual this coalition is. Altman, Amodei, and Hassabis are competing for the same talent, the same compute, and the same customers, and they disagree sharply on questions ranging from open weights to the pace of regulation. On DNA screening, the alignment is easier to explain. The ask targets the wet-lab supply chain, not the AI models themselves. None of these companies sells synthetic DNA, so endorsing mandatory screening costs them nothing operationally while letting them get ahead of a genuinely frightening misuse vector that, if it materialized, would land at their feet.

There is also a liability and timing dimension that is hard to ignore. The letter lands as both Anthropic and OpenAI move toward public markets, with S-1 filings putting their risk disclosures under a microscope. Backing a concrete, narrowly scoped biosecurity guardrail lets the labs demonstrate responsibility to regulators and investors alike, and to point to an external chokepoint, the DNA provider, as a backstop if their models are ever implicated in a real-world bio incident. Supporting a single national standard also heads off the patchwork of conflicting state laws the letter explicitly warns against.

That said, the underlying concern is not manufactured. Independent biosecurity researchers have warned for over a year that current screening can be evaded, including by AI-generated sequence variants engineered to slip past pattern-matching filters, and some argue S. 3741 as written does not fully address that gap. The CEOs' endorsement gives the issue political oxygen it has lacked, even if the technical fix is harder than the headline suggests.

What to watch

The immediate question is whether Commerce Committee leadership moves S. 3741 to markup this session, as the letter urges, or whether the voluntary H.R. 3029 approach gains traction instead. Watch for how the bill handles enforcement and the dangerous-sequence list, who maintains it, how fast it updates, and whether it can keep pace with AI-designed variants. Also watch whether the rare CEO unity holds once the details of compliance costs, recordkeeping mandates, and Commerce rulemaking authority come into focus. Consensus on principle is cheap. Consensus on a statute is the hard part.