JupiteX Get the app
Defence & Security20 Nov 2026 · about 7 min

Why the Next Bioterrorists May Not Make the Same Mistakes as Aum Shinrikyo

The brief

Bioterrorism generally means the deliberate use or threatened use of biological agents to harm people, animals, or plants in pursuit of a political or ideological goal. This definition draws on established usage; the article itself does not give a formal definition. It focuses on terrorist groups’ efforts to develop biological weapons and the risks those efforts may pose. Aum Shinrikyo is the article’s central example. The Japanese cult had money, facilities, and scientifically educated members, yet its biological weapons program repeatedly failed. Its members tried to work with bacterial pathogens and toxins, but faced problems ranging from finding suitable material to making and deploying a weapon. The article highlights those difficulties without treating failure as proof that such efforts are impossible. The article also describes AI as a potential way to ease technical and organizational barriers. Interviews with former Boko Haram members found factions using AI to route difficult questions, share guidance, and learn after unsuccessful operations. The authors argue that past failures should not lead policymakers to dismiss future risks or preparedness concerns.

01

What does bioterrorism mean?

Bioterrorism generally means the deliberate use or threatened use of biological agents to harm people, animals, or plants in pursuit of a political or ideological goal. This definition draws on established usage; the article itself does not give a formal definition. It focuses on terrorist groups’ efforts to develop biological weapons and the risks those efforts may pose.

Aum Shinrikyo is the article’s central example. The Japanese cult had money, facilities, and scientifically educated members, yet its biological weapons program repeatedly failed. Its members tried to work with bacterial pathogens and toxins, but faced problems ranging from finding suitable material to making and deploying a weapon. The article highlights those difficulties without treating failure as proof that such efforts are impossible.

The article also describes AI as a potential way to ease technical and organizational barriers. Interviews with former Boko Haram members found factions using AI to route difficult questions, share guidance, and learn after unsuccessful operations. The authors argue that past failures should not lead policymakers to dismiss future risks or preparedness concerns.

02

What biological weapons did Aum Shinrikyo try to develop, and what went wrong with its efforts?

Aum Shinrikyo pursued biological weapons based mainly on bacterial pathogens and their toxins. The article names Bacillus anthracis as one target and describes the group’s difficulty acquiring or engineering a virulent strain. The broader lesson is that having scientific staff does not mean an organization has the right skills for every step of a weapons program.

Seiichi Endo led the program, but his veterinary education and unfinished graduate study in virology did not provide in-depth knowledge of the specific agents Aum sought to weaponize. The group also struggled with large-scale production and deployment. Under time pressure, members without biology or engineering training improvised equipment. A sprayer they built broke down repeatedly and spouted during attempted attacks.

The article presents these failures as the result of both technical barriers and organizational problems, not proof that future groups cannot succeed. It argues that AI may help extremists access knowledge and learn from failed operations. Policymakers, security analysts, and biosecurity researchers are urged not to treat Aum’s experience as a guarantee of future failure.

03

How many different fields of expertise does the article say a biological weapons program may require, and what are they?

The article says a biological weapons program may require expertise across microbiology, aerobiology, chemistry, statistics, and mechanical and electrical engineering. That is five grouped areas in the article’s wording, or six fields if mechanical and electrical engineering are counted separately. The distinction matters because the work is not one laboratory task.

A program also has to connect different subtasks, from identifying and cultivating suitable material to scaling, purification, weaponization, and deployment. A person may understand one part without being able to solve the others. Aum’s leader, Seiichi Endo, had relevant education, but the article says his background did not cover all the pathogen-specific and operational knowledge the effort demanded. Other members were assigned tasks without biology training, and equipment was improvised without relevant engineering skills.

The article’s broader point is that organizations need both varied expertise and ways to coordinate it effectively. AI could potentially help groups route questions and circulate guidance, as reported among Boko Haram factions. The authors therefore warn against assuming that Aum’s past failure means future groups will fail too.

04

How did secrecy, hierarchy, and internal competition make it harder for Aum to use knowledge effectively?

Aum’s organizational structure made it harder to turn knowledge into sound decisions. Secrecy restricted access to relevant expertise, while hierarchy gave authority to people who lacked the technical background to judge important choices. Internal competition also changed incentives. Together, these dynamics weakened the group’s ability to apply what members knew.

The article describes how members delegated key production and deployment tasks to people with no biology training. Under time pressure and fear of detection, they improvised equipment rather than relying on appropriate skills. When a commercial spraying system was expected to take two months to obtain, Asahara ordered members to build one themselves. The resulting sprayer repeatedly broke down, and its performance was unreliable.

The authors’ point is that technical capability depends on more than recruiting a specialist. Knowledge must reach the right people and guide decisions, and projects need reliable validation before moving toward deployment. The article warns that AI may help extremist groups share knowledge and work through problems, so Aum’s organizational failures should not be treated as a permanent barrier to others.

05

How could AI change the way extremist groups access expertise and learn from failed attempts?

The article argues that AI could help address both technical and organizational hurdles that contributed to Aum Shinrikyo’s failures. In principle, it may help extremists access relevant knowledge, circulate guidance, and improve their ability to work through difficult questions. This matters because Aum’s problems were not only scientific; secrecy, hierarchy, and poor coordination also kept knowledge from guiding decisions.

Interviews with former Boko Haram members, conducted by Juelich, describe factions building specialized AI units. Those units route difficult questions, supplement members’ knowledge, circulate guidance, and help them iterate after unsuccessful operations. The article presents this as evidence that extremist groups are already using AI to support learning and coordination.

The authors do not claim that AI guarantees successful weapons development. They argue that it could reduce some barriers and that past failures by well-resourced groups cannot establish that future groups will fail. Policymakers, security analysts, and biosecurity researchers should therefore take the possibility seriously and consider preparedness, rather than relying on Aum’s record as reassurance.

06

How are Boko Haram factions described in the article as using AI?

The article describes Boko Haram factions as developing specialized AI units to support members’ access to expertise. The reported uses include routing difficult questions, supplementing knowledge, and circulating guidance. The point is that AI may serve as an organizational aid as well as a source of information, helping members connect questions with guidance.

The account is based on interviews Juelich conducted with former Boko Haram members. According to the resulting report, factions also use these units to help members iterate after unsuccessful operations. The article presents this as a way to learn from setbacks, addressing some of the organizational weaknesses that hindered Aum, where knowledge did not reliably guide key choices.

The article does not say that AI guarantees success or removes the scientific challenges of biological weapons. It argues that AI has the potential to ease both technical and organizational barriers and that extremists are already using it. For that reason, the authors caution against treating Aum’s past failures as proof that non-state groups will fail again.

07

Why isn’t having a biologist the same as having all the expertise needed to develop and deploy a biological weapon?

A biologist’s knowledge is not a complete weapons-development skill set. The article explains that biological weapons work involves interdependent steps, including acquiring and identifying material, cultivation, scaling, purification, weaponization, and deployment. Each step can demand different knowledge, and the work must be coordinated across disciplines. That is why a program may require more than one kind of expertise.

Seiichi Endo was appointed head of Aum’s bioweapons program because of related education. He studied veterinary medicine and began, but did not finish, graduate study in virology. Yet Aum focused on bacterial pathogens and toxins, and Endo lacked in-depth pathogen-specific knowledge of agents the group tried to weaponize. His expertise also did not cover large-scale production and deployment. Members without biology or engineering training handled key tasks and improvised equipment.

The article says relevant fields include microbiology, aerobiology, chemistry, statistics, and mechanical and electrical engineering. It also stresses that expertise must be applied effectively: secrecy, hierarchy, and competition can prevent sound decisions. AI may help groups circulate guidance, making past failures an unreliable guarantee of future failure.

This brief was written by AI from the original reporting and checked by other models. Names, figures and quotes come from the source; read it for full context.

Read more in the JupiteX app

Pulse is free. New stories every 4 hours, each one broken into the questions that explain it.

Or read more news on the web