AI Leaders Push Congress for DNA Screening Laws

The people building the world's most powerful AI systems are asking Congress to put up guardrails before their technology makes bioweapons too easy to create.

Sam Altman of OpenAI, Demis Hassabis of Google DeepMind, Dario Amodei of Anthropic, and Mustafa Suleyman of Microsoft AI just signed a public letter supporting new laws that would require DNA synthesis companies to screen every customer and order. The goal is simple: stop someone from mail-ordering the genetic building blocks of a deadly virus.

The timing matters because AI is rapidly lowering the knowledge barrier to creating biological weapons. Right now, you can order custom DNA sequences online from dozens of companies around the world, and not all of them check what you're making or why.

Back in 2017, Canadian researchers spent $100,000 on mail-order DNA to recreate the extinct horsepox virus, proving the concept works. Gene synthesis has only gotten cheaper since then, and AI models can now help design dangerous toxins and pathogens much faster than before.

"AI tools enable a user to very quickly identify where to turn to order sequences that will not be subject to screening," explains Stanford microbiologist David Relman, who also signed the letter. The models can even suggest tweaks to orders that slip past existing screening software.

Many synthesis companies already voluntarily screen for dangerous sequences, but there's no legal requirement. A bipartisan Senate bill introduced this year would change that, making screening mandatory for all providers operating in the US.

The Bright Side

This letter represents something rare: tech leaders asking for regulation before disaster strikes instead of after. The signers include not just AI executives but also DNA synthesis company leaders, biosecurity experts, and national security officials working together on solutions.

Companies like Twist Bioscience have supported formal screening rules for years and already check orders against databases of concerning genetic sequences. Federal guidelines introduced under the Biden administration took a first step by requiring government-funded researchers to use screened DNA sources.

The letter's authors acknowledge that screening isn't perfect, which is why they're calling for AI companies to add their own safety measures. James Diggans of Twist Bioscience puts it simply: "If you have technology that is capable of synthesizing DNA, then you should ensure that it's used responsibly."

Geoff Ralston of the Safe AI Fund believes AI biology models should make it nearly impossible to ask them for help creating weapons. "It should be very difficult, if not impossible, to ask a model to help you do something imminently dangerous," he says.

The collaboration between AI leaders, biotech companies, and policymakers shows an industry willing to accept limits on its own technology when lives are at stake.