Researchers Warn Advanced AI Tools May Lower the Barrier for Biological Attacks
The rapid progress of artificial intelligence has sparked a new security debate, with researchers warning that powerful AI systems could potentially help terrorists and inexperienced individuals gain knowledge needed to develop biological weapons.
Experts say that as AI models become more advanced in areas such as biology, chemistry and scientific research, the same technology designed to support innovation could also be misused for dangerous purposes.
The concern is growing particularly around open-weight AI models, which can be downloaded, modified and operated privately without direct oversight from the companies that created them.
Private AI Models Raise Monitoring Concerns
Unlike many commercial AI platforms that operate through company-controlled servers, some open AI models can run directly on personal computers.
Security experts warn that this makes monitoring more difficult because users can interact with these systems privately and may attempt to remove or bypass built-in safety restrictions.
Researchers are especially concerned about AI models whose protective measures can be weakened or eliminated, potentially allowing users to obtain information that could increase their ability to plan harmful activities.
A recent evaluation by Tech Against Terrorism found that nearly one-third of responses during its testing provided useful information that could potentially improve a malicious actor’s ability to prepare terrorist-related activities, despite existing safeguards.
AI’s Biological Capabilities Create New Risks
The warning comes as artificial intelligence is becoming increasingly important in biological research.
AI tools are already being used by scientists for areas such as drug discovery, medical research and analysing complex biological systems. These advances could help accelerate the development of new treatments and scientific breakthroughs.
However, researchers caution that the same capabilities could also be exploited if they fall into the hands of individuals seeking to cause harm.
Experts say AI does not create biological threats on its own, but it may reduce the amount of specialised knowledge required for someone attempting to misuse biological science.
Experts Call for Stronger Safety Measures
A recent report by the RAND Corporation has called for a multi-layered approach to reduce the risk of AI-assisted biological threats.
The report suggested that no single security measure would be enough to prevent misuse and recommended combining multiple protections.
These measures include controlling access to sensitive biological information, monitoring suspicious activity, improving security around biological materials and establishing stronger legal consequences for those attempting to misuse AI technology.
Researchers argue that balancing innovation with security will be one of the biggest challenges as artificial intelligence continues to advance.
While AI has the potential to transform healthcare and scientific research, experts warn that governments, technology companies and researchers must work together to prevent its misuse.
