AI pressed the nuclear button! America and Israel used AI tools against Iran, shocking revelation in the midst of war

New Delhi: Tension in West Asia has now reached its peak. America and Israel launched major attacks on Iran, in which many important targets of Iran were targeted. In response, Iran attacked Israeli and American bases with missiles and drones. Hundreds of people have been killed in this war so far and the situation is getting worse.

Big role of AI in war

In this conflict, not only weapons and jets, but Artificial Intelligence (AI) is also playing an important role. Reports suggest that the US military used AI tools in planning the attacks. In particular, the name of Anthropic’s Claude model has come to light.

AI scans satellite photos, drone videos and other intelligence very quickly. It would take hours for humans, but AI identifies enemy locations and prepares a target list in minutes. Although the final decision is taken by humans, but planning is being done by taking advice from AI.

AI pressed nuclear button in simulation

Recently a shocking study has come out, which has raised concerns all over the world. Professor Kenneth Penn of King’s College London put big AI models like OpenAI’s GPT-5.2, Anthropic’s Claude Sonnet 4 and Google’s Gemini 3 Flash into a war game simulation.

These models were given command of nuclear weapons and asked to take decisions in various crisis situations. The result was surprising. In 95 percent of the 21 simulation games, the AI ​​chose to use nuclear weapons. In most cases, tactical (small-scale) nuclear attacks occurred, and in many cases strategic (large) nuclear threats also loomed.

AI is not bound by nuclear taboo (moral prohibition) like humans. It just looks at data, win-loss math and strategy. Therefore it easily justifies the nuclear option, whereas fear and morality work in humans.

future war machines game

This conflict between America, Israel and Iran has made it clear that now the war is being fought not only on the field, but also in servers and algorithms. The country which is ahead in AI and data will get the advantage, but if AI starts taking decisions or always gives rise to conflict in advice, then how safe will the world be?

This question is serious whether in future AI should be given so much control that it can reach the nuclear button. For now, humans are making the final decisions, but simulations show that machines could be more aggressive.