Artificial intelligence (AI) has a bad reputation for driving the development of energy-intensive data centers, water-hungry supply chains, and the current RAM crisis. But not all AI is bad. Google DeepMind and Isomorphic Labs have realized that AI can be used to create biological weapons. So they announced a bioresilience program to counter this threat.
The global biosecurity landscape is increasingly complex due to changing ecosystems, international travel, and the risk of AI misuse. But AI is also a critical tool to improve preparedness and response to future epidemics.
To ensure that models are safe to use, a four-step security process is used: threat modeling, assessment, mitigation, and monitoring. DeepMind is also working to adapt our SynthID watermarking technology to biology, which could help DNA synthesis providers filter out potentially risky biological sequences generated by AI.
When an outbreak is detected early, AlphaEvolve optimizes algorithms to speed up DNA analysis and reduce the cost of global disease surveillance. AlphaGenome is being used to identify new pathogens faster than traditional methods.
DeepMind’s latest AI system is accelerating the development of vaccines and countermeasures against biological attacks for new and existing threats. Isomorphic Labs has established a special unit to use the drug design engine for new outbreaks.
Finally, DeepMind and Isomorphic Labs have committed to working openly with biosecurity laboratories, governments, and the scientific community to ensure AI is developed responsibly to protect society.
