Scientists say AI-made bioweapons are a lower risk than many fear

In short: Worries about AI creating a deadly plague are growing, but several scientists say bioweapons are not the most likely path to an AI disaster.

What's going on

Public anxiety about “rogue AI” is rising, and one common fear is that an AI system could help create biological weapons. This includes tools that can write text and answer questions, often called large language models (like a very fast, very broad study helper).

In recent months, AI leaders have pushed for stricter rules around synthetic DNA, which is lab-made genetic material. Researchers have also shown that AI can help design new viral genomes, which are the instruction manuals inside viruses. Anthropic, the company behind the Claude chatbot, also reported attempts to use its system in ways that could support biological weapons work.

Even so, experts quoted by Wired say the biggest barriers are still physical. David Bellamy, a research scientist, argues that finding information is not the main problem. The hard part is getting the gene fragments, equipment, and lab expertise needed to build a virus, test if it infects humans, and see if it spreads between people.

Jason Kelly, CEO of Ginkgo Bioworks, described a project where an AI model ran parts of a lab workflow, but it still could not “take over” a lab because humans control access to materials. Other researchers also say fully autonomous labs are not widely available today, which limits what AI can do on its own.

What to watch

Some experts still worry about bad actors using AI as a helper, not as an independent villain. Suggested safeguards include screening orders at companies that sell synthetic DNA and RNA, improving outbreak detection systems, and using “layered” protections, like multiple locks on the same door.

Source: Wired