Artificial intelligence is beginning to reshape biology and medicine, but scientists say stronger safeguards will be needed as the technology becomes more capable.
A recent experiment by researchers at Stanford University and the Arc Institute showed how AI can already be used to design genetic instructions for viruses. The researchers focused on viruses that infect bacteria and are not capable of infecting humans.
The scientists asked an AI system to generate genetic designs for these viruses. They then produced DNA based on the AI-generated instructions and introduced it into bacteria. Some of the bacteria began producing new viruses, which were able to infect other bacterial cells.
The researchers created and tested 285 designs, of which 16 were successful. The study was first shared last September and was published in the journal Science last month.
The experiment highlights both the potential and the risks of using AI in biological research. AI systems can rapidly explore genetic possibilities that might take researchers much longer to identify through traditional methods. This could eventually help scientists understand diseases, develop medicines and explore new biological treatments.
However, the same capabilities raise concerns about how AI could be misused if biological design tools become widely available without appropriate controls. Even when experiments are conducted using viruses that are harmless to humans, the ability of AI to generate biological instructions demonstrates how quickly the technology is advancing.
Researchers therefore argue that humans must remain closely involved as AI systems become more capable in biology. Scientific oversight, safety checks and responsible access to biological design tools will be increasingly important.
The development also reflects a broader shift in scientific research, where AI is moving beyond analysing existing biological information and beginning to actively assist in designing new biological systems.
As these tools improve, experts say the challenge will be to capture their benefits for medicine and scientific discovery while ensuring that safety measures keep pace with technological progress.
The recent study does not suggest that AI has independently created a human-infecting virus. Instead, it demonstrates that AI can generate functional genetic designs for certain bacterial viruses under controlled research conditions.
Scientists say keeping humans involved in decisions, testing and oversight will be essential as the boundary between artificial intelligence and biological research continues to evolve.
Reader Comments 0