Concerns about AI creating deadly bioweapons are heating up. Yet, many experts argue that the real danger lies in rogue actors, not superintelligent AI. Synthetic DNA laws are being proposed, but the process of creating a virus remains a complex, human-centric task. Even advanced AI can't fully automate the delicate work of lab science.

Despite the alarm, David Bellamy from the Institute of Foundation Models believes that AI's role is more facilitative than transformative. 'AI is a tool, not a threat,' he says, acknowledging that the current bottleneck lies in the hands-on, human-operated labs that stand between a theoretical risk and a real one.

Jason Kelly from Ginkgo Bioworks adds that a superintelligent AI would need to overcome significant obstacles, including the active refusal of lab personnel to assist in such nefarious activities. He argues that the presence of humans acting as checks and balances is a critical deterrent against an AI taking over a lab and deploying bioweapons.

However, there are those who warn of the potential for misuse. Olivia Scharfman from the Institute for Progress highlights the risk of bad actors using AI to commission bioweapons, noting that some fringe groups might be inclined to do so for ideological reasons. She advocates for enhanced protections against all biological threats, including existing pathogens.

Genetic biologist Francois Belloux points out that bioweapons are generally not the most effective method of genocidal violence. 'If you want to kill people,' he says, 'there are much, much, much better ways than to try to engineer some virus or bacterium.' Despite this, the article argues that the focus on AI in bioweapons development is a timely opportunity to strengthen security measures around DNA synthesis and other biotechnologies.