Google DeepMind and Isomorphic Labs outline joint bioresilience strategy using AI models and agents

Google DeepMind and Isomorphic Labs described a joint program focused on bioresilience that combines AI model development with operational partnerships to address infectious disease risks. The effort targets three areas: prevention of model misuse, rapid detection of outbreaks, and accelerated response through therapeutic design.

To prevent misuse, the organizations report using a four-step safety process—threat modeling, evaluations, mitigations, and monitoring—developed with internal biologists, security experts, and external partners. They also note work to adapt SynthID watermarking technology for screening AI-generated biological sequences at DNA synthesis providers.

For detection, the partners highlight AI agents such as AlphaEvolve, which optimizes algorithms used in metagenomic sequencing to reduce cost and speed up pathogen surveillance. They also describe exploration of AlphaGenome and protein function annotation to detect novel pathogen patterns from sequence data more quickly than traditional methods.

In response, the organizations state they are granting trusted researchers access to Google DeepMind’s latest AI systems to accelerate vaccine and countermeasure design, including via AlphaFold. Isomorphic Labs reports establishing a focused unit to rapidly deploy its Drug Design Engine to support governments and non-profits during novel outbreaks, aiming to design medical countermeasures for both natural pandemics and AI-related biological risks.

Over the past 12 months, the partners report advancing more than 15 partnerships with government bodies, biosecurity organizations, and research groups to implement these safeguards and capabilities.