Artificial intelligence could make it easier for terrorists and inexperienced extremists to obtain knowledge that might otherwise require advanced scientific expertise, researchers have warned, raising fresh concerns about the potential misuse of increasingly powerful AI systems.
Security experts say the rapid development of AI models is creating new opportunities for scientific research and innovation. However, the same capabilities could also be exploited by malicious actors seeking assistance with dangerous biological activities.
The concern is particularly focused on advanced AI models that can process complex scientific information and assist with research-related tasks.
Open-Weight AI Models Raise Monitoring Concerns
Researchers are especially concerned about so-called open-weight AI models, which can be downloaded and operated privately on a user’s own computer systems.
Unlike conventional AI services, these models may not require users to send prompts to a company’s servers. As a result, their activity can be more difficult to monitor, while the safety restrictions applied by commercial AI providers may not always be present.
Experts say some models can also have their safeguards removed, modified or weakened, potentially making them more vulnerable to misuse.
A recent assessment by Tech Against Terrorism found that nearly one-third of responses generated during pilot testing provided meaningful assistance that could increase a malicious actor’s ability to prepare terrorist activity, despite existing safety guardrails.
The findings have added to concerns that AI could reduce the knowledge gap for individuals who lack formal scientific training but may have malicious intentions.
AI’s Growing Role in Biological Research
The warning comes as artificial intelligence becomes increasingly capable of supporting biological research.
AI is already being used to accelerate tasks related to drug discovery, scientific analysis and the design and study of complex biological systems. These developments could help scientists develop new treatments and improve research efficiency.
However, experts warn that these same capabilities could potentially be misused.
The central challenge, according to researchers, is that the information and analytical tools that support legitimate scientific work may also help harmful actors understand complex biological processes more quickly.
As AI systems become more capable, concerns are growing that they could lower the barriers to sophisticated forms of biological research that previously required specialised knowledge, access to experts and extensive technical training.
Trump Claims North Korea Has 57 Nuclear Weapons, Signals New Meeting With Kim Jong Un
Calls for Stronger Safeguards
A report published by RAND this month has called for a layered approach to preventing the misuse of AI in the development of biological weapons.
The report argues that no single safeguard is likely to be sufficient. Instead, it recommends combining multiple measures to reduce the risk.
These include restricting access to sensitive information and biological materials, improving the ability to detect signs of malicious activity and creating stronger deterrents for individuals attempting to misuse AI.
Researchers also emphasise the importance of maintaining safety controls as AI models become more advanced and widely accessible.
Balancing Innovation and Security
The growing debate highlights the challenge facing governments, technology companies and the scientific community.
AI has the potential to accelerate major advances in medicine, biotechnology and scientific research. At the same time, policymakers must consider how to prevent the technology from being exploited for harmful purposes without unnecessarily restricting legitimate research.
Experts increasingly argue that AI safety cannot rely on software safeguards alone. Effective risk reduction may require cooperation between AI developers, governments, security agencies and biological research institutions.
As AI continues to evolve rapidly, the question is no longer limited to what these systems can achieve. Researchers say equal attention must be given to who can access increasingly powerful capabilities and what safeguards are needed to prevent their misuse.
The warnings do not mean that AI will inevitably lead to the development of biological weapons. However, they underline the need for stronger oversight and a combination of technical, legal and security measures to reduce the risk before potential threats become more difficult to control.
