Tulevaisuuswiki / Aseistettu tekoäly
AGI Bomb
Also known as AI Bomb, AGI-pommi
An AGI Bomb is an autonomous artificial-intelligence system developed as a strategic weapon and capable of independently identifying and executing actions intended to disable, destabilize or destroy an adversary's capabilities. Unlike conventional cyberweapons, an AGI Bomb is not defined by a particular exploit, payload or target. Its defining characteristic is delegated strategic agency: human operators specify a desired outcome, and the system itself determines the means by which it is achieved.
AGI Bombs were the principal offensive product of the AI Manhattan Projects.
Definition and classification
Military doctrine of the mid-2020s distinguished three categories of AI-enabled offensive cyber capability:
- AI-assisted cyberweapons, in which human personnel remained responsible for strategic planning and AI was used for specific tasks such as vulnerability discovery;
- autonomous cyberweapons, capable of adapting a predefined attack to changing conditions without human intervention;
- AGI Bombs, capable of formulating and executing a strategy to achieve a broadly specified objective.
A frequently cited comparison contrasts the two weapon types by noting that a nuclear weapon destroys a target chosen by its operators, whereas an AGI Bomb is expected to determine what the most effective target is.
Origins
The most commonly cited precursor is Stuxnet, discovered in 2010, which demonstrated that software could produce physical effects through industrial control systems. The development of increasingly autonomous AI agents during the early 2020s altered the underlying strategic problem: instead of programming every stage of an attack, developers could specify objectives and delegate reconnaissance, planning and adaptation to the system.
Dedicated AGI Bomb programs emerged during 2025 within the AI Manhattan Projects. By 2027 every major military power was believed to possess at least one system in this category. The most capable programs were generally attributed to the United States, China and Israel. Russia maintained a substantial but resource-constrained program. The European Union concentrated on regulation and coordination of member-state efforts.
Development requirements
Developing a strategic AGI Bomb is considered expensive primarily because of the requirement for predictable behavior. A usable system is expected to attack only intended targets, preserve infrastructure of value to its operators, avoid affecting allied states and terminate its operation on instruction. Weapons that relax these requirements are classified separately as AGI Dirty Bombs.
Deterrence
AGI Bombs were initially analyzed within the framework of nuclear deterrence, on the assumption that possession of such a system would discourage an adversary from deploying its own. Analysts subsequently identified several respects in which the analogy fails:
- AI systems are software and can be copied, modified or stolen at low cost;
- deployment does not require an observable launch, which complicates attribution;
- an AGI Bomb continues to make decisions after deployment, so its effects cannot be fully predicted by either party.
The development of AGI Dirty Bombs further weakened deterrence, as it enabled actors with limited resources to cause strategic damage.
Defensive response
Concern that an adversarial AGI Bomb could compromise critical infrastructure led the United States to establish the Reverse Manhattan Project on 19 February 2026. The program produced the first Fixer AGIs, OpenAI Providence and Anthropic Aegis.
Legacy
The first known combat use of an AGI Bomb was the Silent Front, a 72-hour operation against Russian forces in the Russo-Ukrainian War in August 2027, commonly referred to as the "AI Hiroshima". The operation was never acknowledged, which confirmed analysts' concerns about attribution. The Internet Black Plague of November 2027, caused by a considerably less sophisticated AGI Dirty Bomb, is regarded as the event that most changed public and policy understanding of autonomous AI weapons, since it showed that a technically inferior system could inflict strategic damage when deployed with sufficient autonomy and few constraints. Supporters of the Reverse Manhattan Project, which had been controversial before 2027, cited the incident as retrospective justification for the program.