AI raises concerns over biological weapons

AI raises concerns over biological weapons

Researchers have warned that advances in artificial intelligence could make it easier for terrorists and inexperienced extremists to obtain information needed to develop biological weapons.

Experts are particularly concerned about sophisticated open-weight AI models that users can download and operate privately. Unlike many cloud-based systems, these models can run locally, potentially giving malicious actors access to powerful AI assistance without having to rely on an external company’s servers.

Researchers say such systems could lower the level of scientific expertise required to plan complex biological attacks. Individuals who previously lacked the technical knowledge needed to understand biological processes could potentially use AI to obtain guidance, analyze information and overcome technical challenges.

Another major concern involves the ability to remove or weaken built-in safety restrictions. Because users can access and modify some open-weight models, safeguards designed to prevent harmful responses may become less effective. Local operation can also make monitoring more difficult because prompts and interactions may never reach the provider’s servers.

Recent research by the Center for a New American Security found that early experiments with AI systems produced concerning results, with roughly one-third of responses offering meaningful assistance that could potentially improve a malicious actor’s ability to prepare for terrorist activity.

At the same time, researchers stress that AI is bringing major benefits to legitimate scientific work. The technology is accelerating biological research, drug discovery and the design and analysis of biological systems. However, the same capabilities that support medical and scientific breakthroughs could also create new risks if malicious actors exploit them.

A RAND report released this month called for a multilayered strategy to address the growing threat. The report recommends strengthening controls over sensitive biological information and materials while improving systems that can identify suspicious activities.

It also calls for stronger safeguards against the misuse of advanced AI and effective legal consequences for individuals who deliberately use the technology to support terrorism or other dangerous activities.

Researchers emphasize that AI itself is not the sole source of the threat. Instead, they say governments, technology companies and the scientific community need to work together to ensure that advances in AI and biotechnology continue to benefit society without giving malicious actors easier access to dangerous capabilities.