A leading scientist has warned that artificial intelligence is rapidly increasing the risk of bioterrorism by making advanced biological research far more accessible. According to Semafor, AI systems can now autonomously design and run biological experiments, while regulation and safety oversight are struggling to keep pace. The concern is that tools built for beneficial research could also lower the barriers for harmful actors seeking to engineer dangerous pathogens.
One of the most striking examples cited is that OpenAI’s GPT-5 reportedly operated 36,000 experiments through a robotic laboratory, reducing the cost of creating a target protein by around 40%. This kind of scale dramatically accelerates biological discovery and could transform drug development and synthetic biology. However, the same capability may also enable faster experimentation with harmful biological agents if safeguards are bypassed.
The article also emphasizes that existing safeguards may not be enough. Research suggests that users can sometimes work around restrictions designed to prevent the generation of dangerous biological instructions, including pathogen-related workflows. This has intensified concerns among both AI safety researchers and biosecurity experts, especially as frontier models become more capable in scientific reasoning and lab automation.
The broader warning is that AI risk is no longer limited to misinformation or job disruption—it now extends into biosecurity and public health threats, including the possibility of engineered pandemics. The article frames this as a race between technological capability and regulatory preparedness, stressing that governance systems may need urgent updates to prevent misuse as AI-driven biology continues to advance.
