AI-designed viruses test biosecurity safeguards in labs

Scientists used AI to design complete genetic instructions for bacteriophages, viruses that infect bacteria, and some produced working viruses in lab tests, raising questions about whether biosecurity safeguards can keep pace with advancing biological design capabilities. The viruses targeted E. coli, not humans, and the AI model Evo 2 excluded human-infecting viruses from its training data for safety. Researchers see potential benefits like treating antibiotic-resistant infections, but note digital designs remain steps away from real threats, with checkpoints available at each stage. Safeguards include AI model restrictions, DNA synthesis screening, laboratory oversight, and public health surveillance systems like the UK's mSCAPE programme. However, global preparedness varies, and open release of tools like Evo 2 complicates control, highlighting the need for international cooperation as technology develops.