North Korea says it tested an AI-guided torpedo

By Splitfeed EditorialLast updated Oct 17, 2026, 05:00 UTC

On October 16 and 17, 2026, North Korea's state media outlet KCNA announced that the country had successfully conducted an underwater test of a new AI-guided torpedo. The Times of India identified the weapon as the Zam-2.0, describing it as a high-speed, precision-guided system. North Korea framed the test as a meaningful step forward in its AI weapons programme, and the announcement was picked up by AP News, The Washington Post, the South China Morning Post, and several other outlets. No independent verification of the test or its claimed results has been produced.

The announcement matters because it adds a new dimension to an already fraught debate about AI applications in military systems. If the capability is genuine, it would represent a rogue state deploying AI guidance in an offensive underwater weapon, a category that carries significant implications for naval security and arms control. The core dispute is whether the claim reflects a real technical advance or follows a pattern of North Korean military theatre. In 2023, Pyongyang unveiled what it called a tactical nuclear attack submarine, which outside analysts identified as a rebuilt diesel boat that had never been seen firing anything until Pyongyang announced cruise missile launches from it. That history shapes how observers are reading this latest announcement.

The anti-AI camp has split between alarm and skepticism. Accounts like FACTSmatter warned that rogue states such as North Korea, Iran, and Russia could use AI agents to attack Western infrastructure and trigger economic chaos. Female Arthur Morgan pointed out that North Korea, Iran, and the Houthis are all engaging with AI capabilities while some observers dismiss the threat as implausible. The account Expatriot took a more sardonic view, suggesting North Korea's presentation was itself a form of propaganda. The middle ground, represented by accounts like Kraken and The Defence Blog, treats the claim cautiously, noting the unverified nature of the AI guidance assertion and pointing to the 2023 submarine episode as a reason to reserve judgment. The pro-AI camp has not reacted publicly to this story; given the subject matter, that camp would typically focus on the broader argument that AI advances are inevitable and that restricting them in Western democracies does nothing to slow adoption by adversarial states.

The argument will remain unresolved without independent technical analysis of the Zam-2.0 or satellite and signals intelligence that corroborates the test. Any formal assessment from South Korean or American defence agencies, or a published technical review of the weapon's actual guidance system, would be the clearest way to distinguish a genuine capability from a staged announcement.