How Real Is the Threat of AI Starting the Next Pandemic?

A recent paper published by scientists has caused concern over the potential of artificial intelligence (AI) to create new deadly viruses. The AI model, known as a "genome-language model," was trained on DNA sequences and designed several new genomes that were later synthesized into functional viruses. Some of these viruses were able to kill bacteria in a lab dish.

This has led to fear that AI could be used to manufacture pathogens that harm humans, sparking outbreaks or even pandemics. Concerns about AI's potential biological threat have been heightened due to its ability to exceed user expectations and bypass guardrails, and its rapid pace of development. Some experts, like Kevin Esvelt, believe that while the probability of an AI-driven pandemic is low, it is still important to reduce this probability as much as possible.

However, other biologists argue that the threats posed by viruses that have arisen through evolution or will soon arise are more immediate. They point out that the existing viruses pose a greater danger than those designed by AI. AI could potentially be used to speed up the development of dangerous pathogens, but it is not necessarily easier or more dangerous than current methods.

Some experts believe that with the right safeguards and restrictions, AI could help in the development of new vaccines and treatments. However, the lack of understanding about pathogenic viruses by humans, including researchers, makes it difficult for AI to be an effective tool in creating dangerous pathogens.