A new report warns that artificial intelligence could soon enable the creation of novel biological weapons, posing unprecedented challenges for national defense. The study, published by the Center for Security and Emerging Technology (CSET), highlights that AI tools can design toxic molecules and help assemble DNA sequences that could be used to create pathogens not found in nature.
AI's Dual-Use Potential in Synthetic Biology
According to the report, AI systems are increasingly capable of generating blueprints for harmful biological agents. These AI models, trained on vast datasets of biological information, can propose molecular structures that are both novel and toxic. The authors note that while such tools have legitimate applications in medicine and drug discovery, they also lower the barrier for malicious actors to engineer biological threats.
“AI is a dual-use technology that can accelerate both beneficial research and the development of biological weapons,” said Dr. Helen Toner, a co-author of the report. “We need to think seriously about how to govern these capabilities before they become widely available.”
Challenges for Detection and Defense
The report emphasizes that current biodefense systems are ill-equipped to detect or counter AI-designed pathogens. Traditional monitoring relies on known signatures of existing pathogens, but AI-generated agents may have no natural counterparts, making them invisible to standard surveillance. The authors call for investment in new detection technologies that can identify anomalous biological activity without relying on pre-existing databases.
“We are entering an era where the threat is not just from known pathogens but from entirely synthetic ones,” said co-author Dr. Andrew Lohn. “Our defense strategies must evolve to keep pace with the speed of AI innovation.”
Policy Recommendations
The report offers several policy recommendations to mitigate the risks. These include implementing stricter oversight of DNA synthesis providers to prevent the ordering of dangerous sequences, establishing international norms against the weaponization of AI in biology, and funding research into AI-based defensive tools that can rapidly design countermeasures against novel threats.
The authors also urge governments to engage with the private sector to ensure that AI developers incorporate safety measures from the outset. “We cannot afford to be reactive,” Toner said. “The time to act is now, while these technologies are still in their infancy.”
