Google DeepMind Launches Bioresilience Program for AI-Driven Biosecurity

Google DeepMind announced Thursday a new bioresilience program designed to deploy artificial intelligence capabilities for biological threat prevention, detection, and response. The initiative represents a strategic bet that the same frontier AI systems raising biosecurity concerns can become essential tools for defending against those risks.

Why it matters

As AI models grow more capable of predicting and manipulating biological systems, the technology industry faces mounting pressure to demonstrate that these tools strengthen rather than undermine global biosecurity. Google's program signals a shift from treating biology purely as an AI safety problem to positioning frontier AI as critical infrastructure for pandemic preparedness.

Program scope and partnerships

The bioresilience initiative, launched in collaboration with Google's Isomorphic Labs, focuses on three core areas: improving pathogen surveillance systems, accelerating vaccine and therapeutic design, and strengthening outbreak response capabilities. According to details first reported by Axios, Google has established more than 15 partnerships over the past year with government agencies, biosecurity organizations, and research institutions.

The company is expanding access to its AI models and agents for what it classifies as trusted partners—researchers, government entities, and biosecurity specialists. Under Google's safety framework, these deployments qualify as "low-risk" restricted releases rather than public launches, allowing the company to maintain tighter controls over who uses the technology and how.

Safety thresholds and testing protocols

Helen King, Google DeepMind's vice president of responsibility, told Axios the company has established clear capability thresholds that would trigger deployment restrictions. "If we were to find that we were reaching a critical capability level and we didn't have the appropriate mitigations, then we would not be launching," King said. She noted that ongoing evaluations conducted throughout development indicate those thresholds have not yet been reached.

Owen Larter, Google DeepMind's senior director of frontier AI policy and global affairs, emphasized that major AI labs have reached consensus on the need for rigorous testing before model releases. "Everyone agrees we can't get this wrong," King said of AI's biological capabilities.

Regulatory alignment

The biosecurity program aligns with broader governance proposals from DeepMind CEO Demis Hassabis, who called earlier this week for establishing a standards body for frontier AI. In an interview with Axios, Hassabis advocated for a "systematic" regulatory approach funded by industry, staffed by technical experts, and accountable to the U.S. government. Larter expressed optimism that the Trump administration's executive order process could produce a "durable framework" for testing frontier AI models before deployment.

Track record in biology

Google has established significant credibility in applying AI to biological research. AlphaFold, the company's protein structure prediction system, earned Hassabis a share of the 2024 Nobel Prize in Chemistry for helping scientists understand diseases and develop new drugs. Isomorphic Labs is now commercializing that expertise for drug discovery applications.

The bioresilience program details were first reported by Axios.