I was sitting across from a printed letter when two signatures caught my eye: Sam Altman’s and Dario Amodei’s. For a moment the room felt smaller, as if a bright problem had been forced into a corner. The sight hit like a loaded gun on a conveyor belt—unexpected, dangerous, and oddly specific.
I’ll speak plainly: you and I both know that AI can accelerate discovery. What worries people now is what happens when that speed meets biology and loosened controls. Below I’ll walk you through who signed, what they asked for, and why lawmakers could be the deciding factor.
Two rival CEOs put ink on the same page
Sam Altman of OpenAI and Dario Amodei of Anthropic signed an open letter alongside leaders from DeepMind and Meta.
That’s not a PR stunt. When Demis Hassabis and Alexandr Wang — the faces of Google DeepMind and Meta AI — add their names, the signal is clear: a broad swath of industry sees a new risk. The letter isn’t a manifesto about AI per se; it’s a targeted ask to policymakers: require DNA synthesis suppliers to screen orders and verify customers before sending out synthetic nucleic acids.
A short story about a shipping label
A DNA order arrives at a lab that hasn’t been vetted.
The letter asks companies that can gatekeep synthetic nucleic acid to check sequences for “sequences of concern” and to log orders so investigators can follow leads. It suggests that traceability can deter misuse. OpenAI, Anthropic and dozens of scientists and policy experts argue this isn’t theoretical—AI’s pace could erode the knowledge barriers that once made biological weapons harder to build.
Can AI design biological weapons?
Yes and no. AI models can generate instructions and propose modifications to sequences, which lowers the technical hurdle for bad actors. But presence of capability doesn’t equal inevitability: governance, lab practices at companies such as Twist Bioscience and Ginkgo Bioworks, and supply-chain checks can still raise the cost of misuse. The letter’s point is that if we let automation and anonymity combine, the balance shifts.
A Friday memo to policymakers
The letter was organized by two think tanks with different political flavors: the Institute for Progress and the Foundation for American Innovation.
Wired reports and Politico covered the effort; the choice of organizers matters because it frames the ask as cross-aisle rather than partisan. OpenAI has recently pushed a federal model-vetting plan in a white paper that goes beyond a recent executive order from President Trump, and Altman even met Bernie Sanders. Those moves read like reputational insurance: companies want to be seen as partners in regulation rather than adversaries.
What did Sam Altman and Dario Amodei sign?
They signed an open letter urging laws that would require DNA synthesis providers to screen sequences and verify customer legitimacy, and to create records that investigators can access. The signatories include people from industry, academia, and policy—names meant to add authority and practical weight to the request.
A lab anecdote about a missed red flag
A technician found a strange order but no paperwork to explain it.
Real labs have horror stories about ambiguous orders and blurred accountability. The letter wants to make those stories rarer by formalizing checks. The hoped-for result is that AI won’t become an accelerant for biological harm. Right now the safeguards resemble a leaky dam: they hold back most water, but a few breaches could flood a valley.
How would DNA synthesis screening stop misuse?
Screening looks for known hazardous sequences and flags suspicious requests, while customer legitimacy checks try to verify that the requester has a legitimate research need. Providers like Thermo Fisher and smaller gene synthesis firms already run variations of this process. The letter asks for legal requirements so that screening and recordkeeping aren’t optional, arguing that making actions traceable changes behavior.
There’s a tension here you should notice: industry leaders want guardrails that keep innovation moving, while national security experts want durable walls that slow dangerous work. I don’t pretend the balance is easy, but I do think the consensus signaled by Altman and Amodei forces a question onto lawmakers’ desks.
If CEOs, labs, and scientists can agree on scanning and traceability, will Congress act fast enough to match the speed of AI and biology?