The rapid development and wider availability of artificial intelligence is generating a new security concern, with emerging research warning that advanced AI systems could potentially make sophisticated biological knowledge more accessible to terrorists and inexperienced extremists.
According to recent research by Tech Against Terrorism, artificial intelligence may make it easier for terrorist actors and individuals with limited technical expertise to obtain information relevant to the development of biological weapons, potentially lowering some of the knowledge barriers that have historically made complex biological attacks difficult to plan.
The findings have intensified concerns over the misuse of increasingly capable AI systems, particularly advanced open-weight models that can be downloaded and operated privately rather than exclusively through centrally managed online platforms.
Researchers have highlighted particular concerns surrounding AI models that can be run locally. Unlike systems operating entirely through a company’s online infrastructure, locally deployed models may process user inputs without those interactions necessarily reaching the model developer’s servers.
This creates an additional challenge for authorities and security researchers attempting to identify suspicious activity. The ability to operate sophisticated AI systems privately may also create opportunities for malicious users to circumvent safeguards that are designed to prevent the generation of harmful or dangerous information.
Researchers warn that safety mechanisms incorporated into AI systems could potentially be weakened, bypassed or removed when models are downloaded and modified for private use. This raises a fundamental security challenge: as AI capabilities become increasingly decentralized, traditional methods of monitoring and controlling potentially harmful uses may become more difficult to apply.
According to the research cited in the report, initial testing found that approximately one-third of responses provided meaningful assistance that could potentially enhance a malicious individual’s ability to prepare for terrorist activity. The finding does not mean that AI systems automatically enable biological attacks, nor does it establish that every such response would translate into real-world capability. However, researchers consider the results significant because they indicate that AI could potentially reduce information barriers for individuals seeking to misuse advanced scientific knowledge.
The concern is particularly serious in the biological domain, where complex scientific expertise, specialized knowledge and access to sensitive materials have traditionally represented substantial barriers.
The emergence of highly capable AI systems has changed the information environment dramatically. Individuals without formal scientific training can increasingly use AI tools to understand complex technical concepts, navigate specialized literature and obtain structured explanations. While these capabilities offer enormous benefits to legitimate research, medicine, education and innovation, the same accessibility can create opportunities for malicious exploitation. The concern is therefore not simply about the existence of dangerous information. It is about the possibility that AI could make sophisticated information easier to locate, interpret and apply.
For counter-terrorism authorities, this represents a rapidly evolving challenge requiring cooperation between governments, technology companies, researchers and international security institutions. Researchers have also raised concerns about the difficulty of monitoring locally operated AI models.
When models operate privately on a user’s own hardware, conventional monitoring mechanisms may have limited visibility into how the system is being used. This could make it more difficult to identify attempts to obtain dangerous assistance or recognize patterns associated with malicious activity. The decentralization of AI capabilities therefore presents a challenge similar to other dual-use technologies: tools developed for legitimate purposes can potentially be repurposed by individuals seeking to cause harm.
Authorities face the difficult task of addressing these risks without unnecessarily restricting legitimate scientific research, technological development or responsible AI use.
The research emphasizes the importance of strengthening safeguards around sensitive information and biological materials while improving mechanisms for identifying suspicious activity. It also calls for effective deterrence and appropriate legal consequences for individuals who deliberately attempt to exploit AI systems for terrorist or other malicious purposes. A comprehensive response will require more than technological restrictions alone. Governments and technology companies will need to strengthen risk assessments, improve safety testing, establish effective reporting mechanisms and develop safeguards capable of adapting to rapidly evolving AI capabilities.
At the same time, international cooperation will be essential because AI systems, digital infrastructure and terrorist networks operate across national boundaries.
The emergence of AI-assisted malicious activity represents a potentially important evolution in the terrorism landscape. Terrorist organizations have historically sought to exploit emerging technologies to enhance recruitment, propaganda, communication, financing and operational planning. The increasing sophistication of AI introduces another potential dimension: access to highly specialized knowledge.
The challenge is particularly complex because the same AI technologies that can accelerate scientific discovery and improve human welfare may also be targeted for misuse. The objective for governments and technology developers must therefore be to preserve the enormous benefits of AI while ensuring that malicious actors cannot easily convert those capabilities into tools for mass harm.
The latest findings underscore the importance of addressing the problem before AI-enabled capabilities become widely integrated into terrorist planning. Preventive measures should focus on securing sensitive information, strengthening AI safety systems, improving detection of malicious use, protecting biological materials and enhancing cooperation between technology companies, law-enforcement agencies and scientific institutions.
Clear legal and regulatory frameworks will also be necessary to establish accountability for deliberate attempts to exploit AI for terrorism or other catastrophic activities. The evolving threat landscape demands continuous reassessment. AI capabilities are developing rapidly, and safeguards that are effective today may not necessarily remain sufficient tomorrow.
