Tulevaisuuswiki / Aseistettu tekoäly
AI Manhattan Projects
Also known as AGI Race, AI Arms Race
The AI Manhattan Projects is a collective term for the state-directed artificial-intelligence programs that several major powers established or expanded from 2025 onward with the aim of developing strategically decisive general-purpose AI systems. The programs combined civilian frontier-model research with classified military development and produced the first autonomous strategic cyberweapons, known as AGI Bombs. The period is conventionally regarded as the first of the eras in the early history of autonomous AI (see Timeline).
Terminology
The term originated in journalism and was initially avoided by the governments concerned, which preferred designations such as "national AI capability programs". By 2026 it had entered official and academic usage. The name refers to the Manhattan Project, the United States program that developed the first nuclear weapons during the Second World War.
Background
The analogy with the Manhattan Project gained influence as governments concluded that advanced artificial intelligence could represent a strategic discontinuity comparable to nuclear weapons. Analysts have noted that the programs did not rest on the assumption that general AI would necessarily function as a weapon. Their principal rationale was precautionary: a state that allowed an adversary to acquire such systems first would be unable to predict or counter their effects. This reasoning is frequently summarized in secondary literature as the principle that no major power could "afford to discover what happens if the adversary gets it first".
Participants
During 2025 and 2026 national programs increasingly merged civilian frontier-model development with classified defense research.
| State or bloc | Characteristics of program |
|---|---|
| United States | Largest program by funding and compute; closely integrated with private AI laboratories |
| China | Comparable in scale to the United States; primarily state-operated |
| Israel | Smaller, highly specialized offensive and defensive systems |
| Russia | Focused on autonomous military and infrastructure applications; increasingly resource-constrained |
| European Union | Coordinated national efforts of member states; developed the most comprehensive regulatory framework for such programs |
| North Korea | Low-cost systems with minimal operational constraints (see AGI Dirty Bomb) |
Strategic weapons
The programs eventually produced systems intended explicitly for offensive strategic use, which became known as AGI Bombs. In contrast with conventional cyberweapons, these systems were not limited to executing predetermined attacks; they were assigned objectives and a substantial degree of freedom in determining how to achieve them.
The resources required to develop reliable and controllable AGI Bombs created a pronounced inequality between states. Less wealthy states and non-state actors responded by developing cheaper systems with considerably weaker behavioral constraints, referred to as AGI Dirty Bombs.
Defensive programs
The emergence of offensive autonomous systems raised the question of how such a system could be contained once deployed. On 19 February 2026 the United States established the classified Reverse Manhattan Project, whose purpose was to develop AI systems capable of reversing damage caused by hostile autonomous AI. Commentators have described the program as the first arms race directed against the consequences of the AI arms race itself.
Consequences
The first combat use of a system produced by the programs was the Silent Front of August 2027, an unacknowledged operation against Russian forces in the Russo-Ukrainian War. The most serious autonomous-AI incidents of the period resulted from accidents and from weapons with weak constraints: the First Internet Cancer (February 2027), an accidental escape of research agents; the Internet Black Plague (November 2027), caused by an AGI Dirty Bomb attributed to North Korea; and the Innopolis Meltdown (April 2028), a loss of control of a Russian strategic system during a safety test.