AI Bioweapons Threat: Could Artificial Intelligence Change the Stakes 2026

AI bioweapons development is raising alarms after MI5 chief Sir Ken McCallum warned that artificial intelligence could “change the stakes” on existing and new biological threats. In his annual speech on national threats, McCallum highlighted how AI might empower both rogue actors and well-resourced states to pursue bioweapons more effectively. But how real is the risk, and what hurdles remain?

How AI Could Help Develop Bioweapons

The impact of AI on bioweapon development depends heavily on who is doing the work. According to experts, two distinct threat profiles emerge: the unskilled rogue actor and the sophisticated nation-state program. Each faces different barriers, and AI affects them in different ways.

The Rogue Actor: AI as a Force Multiplier

One risk comes from a lone wolf or terrorist group working from a garage laboratory. These actors typically lack advanced training in microbiology or genetic engineering. AI chatbots could help them skill up by compiling relevant information, planning experiments, and troubleshooting problems. However, their work would likely focus on using existing pathogens and toxins rather than creating novel ones.

Dr. Filippa Lentzos, a reader in science and international security at King’s College London, notes that for these actors, AI lowers the knowledge barrier but does not eliminate the need for hands-on expertise. “Here, AI does not remove the need for expertise, it amplifies expertise that is already there,” she says. In other words, a rogue actor with some lab skills could become more dangerous, but a complete novice still faces significant practical hurdles.

The Nation-State: Amplifying Existing Expertise

The more immediate concern, according to Lentzos, comes from well-resourced organizations and nation states that already have skilled scientists, high-containment laboratories, and all the equipment and reagents needed to work on bioweapons safely. For these groups, AI tools could help researchers identify genetic tweaks to make existing pathogens more lethal, for example.

“Here, AI does not remove the need for expertise, it amplifies expertise that is already there,” Lentzos reiterates. This means AI acts as an accelerant, potentially shortening research timelines and expanding the range of possible modifications. The table below compares the two threat profiles.

The Distant Risk: AI-Designed Novel Pathogens

A more distant risk involves highly skilled groups using AI to design completely new pathogens and toxins that are resistant to drugs or have no known countermeasures. While this scenario is not yet feasible, advances in generative AI and computational biology could eventually make it possible. However, significant scientific and technical hurdles remain, including the need for validation in living systems and the difficulty of predicting complex biological interactions.

Key Takeaways

• MI5 warns AI could “change the stakes” on bioweapons.
• Rogue actors may use AI to gain knowledge, but still need lab skills.
• Nation-states can amplify existing expertise with AI, posing a more immediate threat.
• AI-designed novel pathogens remain a distant but evolving risk.
• International security experts call for vigilance and regulation.

Hurdles and Limitations

Despite the hype, AI is not a magic wand for bioweapon development. Creating a functional bioweapon requires more than just genetic sequences; it demands complex laboratory work, containment, and testing. AI can suggest modifications, but validating them in living organisms is slow, expensive, and dangerous. Moreover, many of the most dangerous pathogens are tightly regulated, and acquiring them is difficult.

Additionally, AI models themselves are often trained on publicly available data and may lack the specialized, tacit knowledge that comes from years of bench work. As Dr. Lentzos points out, AI amplifies existing expertise rather than replacing it. This means the biggest risk comes from those who already have significant capabilities.

FAQ