A growing governance gap at the intersection of artificial intelligence and biotechnology. AI is accelerating biological research by helping scientists analyze biological data, design molecules and proteins, and explore biological systems at a speed that was previously difficult to achieve. The problem is that AI and biotechnology have traditionally been governed as separate areas, while their convergence creates capabilities—and risks—that do not fit neatly into either regulatory framework.
This convergence matters because AI can reduce the time and expertise required to perform parts of biological discovery and engineering. A system that can reason over biological sequences or structures can potentially accelerate beneficial work in medicine, agriculture and environmental science, but the same capabilities can introduce dual-use concerns. Biotechnology already has a long history of requiring safety and ethical oversight because biological systems can be deliberately manipulated; adding increasingly capable AI can make the pace and accessibility of that manipulation change substantially. RAND researchers similarly argue that the rapid progress of biotechnology is increasingly outpacing governments' ability to develop effective regulatory responses.
The difficult part is that existing rules often regulate the technology rather than the combined capability. AI regulations may focus on models, data or automated decision-making, while biotechnology rules focus on laboratories, organisms, medical products or genetic manipulation. An AI system that assists biological design can sit between those categories. This creates questions about who should be responsible—the AI developer, model provider, laboratory, researcher, data provider or end user—and at what stage safeguards should be applied. Recent research on AI-biotech governance similarly argues for approaches that account for the interaction between the two technologies rather than treating them independently.
The broader message is that regulation needs to become more adaptive and capability-based. Waiting for a specific technology or biological application to become widespread before creating rules risks leaving regulators permanently behind the innovation cycle. At the same time, overly broad restrictions could interfere with legitimate scientific research. A more practical approach would combine risk assessment, laboratory controls, AI-system safeguards, monitoring, accountability and international coordination.
