AI leaders call for tougher protections against AI-powered bioweapons

Some of the biggest competitors in the AI ​​industry have set their goals a lot, a lot Injustices Aside for a common reason: to make it harder for people to use their technology to develop biological weapons. in Open letter As for US lawmakers, technology leaders are pressing Congress to enact rules that close what they say is a troubling biosecurity gap that could help spark a global pandemic.

Anthropic’s Dario Amodei, OpenAI’s Sam Altman, and Microsoft’s Mustafa Soliman were among the signatories who urged US lawmakers to require companies that sell DNA and synthetic RNA — genetic material that can be ordered online and assembled in a lab — to screen purchases for sequences that could be used to make dangerous pathogens. The fear is that AI tools could make it easier to design potentially dangerous sequences, order them from manufacturers, and use them in ways that previously required more specialized expertise.

Other signatories include Meta AI chief Alexander Wang, and Google DeepMind’s Demis Hassabis, who won the 2024 Nobel Prize in Chemistry for his work on AI-based protein prediction. The letter was also signed by prominent scientists, national security and policy experts, and executives from biotechnology companies including Twist Bioscience and Ansa Biotechnologies, both major vendors of synthetic genetic materials. It was the message It is said It is organized by two think tanks: the Foundation for American Innovation and the Institute for Progress.

Scientists have long warned that advances in synthetic biology could make it easier for scientists to engineer dangerous or even dangerous organisms Reviving long-dead pathogens – Work that can cause destruction if misused, mishandled or released by accident. But this power has remained largely in the hands of skilled scientists with access to sophisticated laboratories, equipment, and resources. The concern now is that as biological tools become less expensive and more accessible, and AI models become more powerful, the barriers to their misuse are beginning to crumble. Experts also warn that AI could help produce other things Threats such as chemical weapons.

While the letter acknowledges that many of the largest providers of synthetic DNA and RNA are already screening applications, this is done on a voluntary rather than mandatory basis. Detailed records must also be kept of any orders, the letter says, in order to track down any threat that evades initial screening.

“Given the pace at which the underlying technology is changing, we believe the need is urgent,” the letter said. “It is a rare moment of agreement between stakeholders who are often at odds. We hope policymakers will confront it with decisive action.”