loading
loading
Iran and Houthi rebels used Anthropic's Claude AI to target US warships and build hypersonic missiles — Houthi rebels also used the bot to code ballistic missile guidance systems 'Great Satan's' AI comes in handy. Iran's spiritual leaders tend to call the U.S. the Great Satan to express their spite, but it turns out that its military, surveillance, propaganda, and even allied Houthis are eager to use American-built AI technology to target the U.S. Navy and develop weapons, surveillance, and propaganda, Anthropic's September 2026 threat report revealed. Arguably, one of Anthropic's most remarkable findings is that an Iran-linked threat actor used an American AI model, Claude, to support military reconnaissance and develop targeting recommendations against U.S. naval forces in the Middle East. The perpetrator combined publicly available ship and aircraft transponder identifiers with commercial satellite imagery and information on U.S. naval movements, and even extracted the names of U.S. military personnel from captions of publicly available military photographs. It also researched potential vulnerabilities in communications equipment used aboard ships, including known flaws affecting Cobham Sailor VSAT terminals, Cisco communications equipment, and Schneider Electric EcoStruxure systems. Anthropic said it banned the account, introduced additional detection mechanisms, and shared its findings with government authorities. Another striking case involved a cell in northern Yemen controlled by Houthis (which are in turn controlled by Iran) that used Claude Code to support three weapons programs: a guided rocket that uses a phone-class flight computer that assists terminal guidance, a multistage ballistic missile targeting a range of more than 2,000 km, and an R2000 missile family that included a hypersonic glide vehicle variant. The group used Claude to develop guidance, navigation, and control software; integrate an open-source autopilot with a phone-class flight computer; write control and position-estimation code; tune parameters; build firmware; and even run flight simulations. Essentially, the group used multiple Claude instances instead of a group of software engineers for coding, code review, research, and simulation. While Houthis are technically not Iranians, they can certainly share their research and development results with their allies and potentially use Iran's industrial capacity to build their weapons. In addition to building targeting recommendations against American naval forces as well as speeding up the development of weapons, Iran used Claude for surveillance tools. One Iran security-linked unit used Claude to analyze 155,216 tweets to profile, identify, and surveil 6,388 opposition individuals in a single year. Another group used the model as an engineering pipeline to develop domestic tracking tools, including the production-deployed "al-Najm al-thāqib" Firefox extension designed to mass-harvest user identities across major social platforms. While Anthropic has banned 16 Claude accounts associated with Iranian paramilitary and domestic security agencies, that does not mean it has banned all of them. Iran-linked actors and Houthis are not the only entities using Anthropic's AI technologies for weapon development. China and Russia are also actively using Claude for their military programs. Get Tom's Hardware's best news and in-depth reviews, straight to your inbox. Follow Tom's Hardware on Google News, or add us as a preferred source, to get our latest news, analysis, & reviews in your feeds. Anton Shilov is a contributing writer at Tom’s Hardware. Over the past couple of decades, he has covered everything from CPUs and GPUs to supercomputers and from modern process technologies and latest fab tools to high-tech industry trends. - FunSurfer Reply Your support will probably end when one of those weapons will hit you...usertests said:I support the use of AI by anyone to create new and exciting weapons. - King_V Reply What if it's new but not exciting? Would you be opposed?usertests said:I support the use of AI by anyone to create new and exciting weapons. - King_V Reply Logically, being no longer alive precludes one from being excited about, or, really, having ANY feelings about, well, anything, really.FunSurfer said:Your support will probably end when one of those weapons will hit you... - usertests Reply I will become a Force Ghost and haunt the survivors.FunSurfer said:Your support will probably end when one of those weapons will hit you... It will have only my tepid endorsement.King_V said:What if it's new but not exciting? Would you be opposed? - me-three The upshot of this is that the best Chinese models aren't far behind the US frontier models, especially in coding. So even if bad (state) actors are entirely banned from US models, it only means that it would only take those engineers a bit more effort to accomplish the same thing with DeepSeek/Kimi K3.Reply The US models may be the best, but the others are "good enough" for 95% of the tasks. In 3-6 months, the bar will have moved, and while the best may still be in front, "good enough" will be where "best" is now. This is why efforts to stop or slow down AI are doomed to fail. Every nation is on a Manhattan's Project to build the best AI, or learn how to best use it, because "national security" is literally, not just rhetorically, at stake. This brings to mind the recent "AI will kill all humankind" blow-up. It's enlightening to understand that AI doomerism is mainly a province of American concerns, not China's. Watch below to learn why, fEGUlvR9gvYView: https://www.youtube.com/watch?v=fEGUlvR9gvY - bit_user Reply Obviously, there are ways to be negatively affected by a weapon without being killed by it. In the case of nukes, the mere threat of their use can affect quite a lot of things, and we've only seen what the world looks like when a small number of nation states have them. A scenario that's popular in works of fiction is a city being held for ransom by a nuke or biological weapon.King_V said:Logically, being no longer alive precludes one from being excited about, or, really, having ANY feelings about, well, anything, really. Weapons in the physical world are only one possible avenue. Cyber weapons are another example, and one where it seems AI has already demonstrated quite substantial prowess. So, this is one of the scenarios where AI can destroy the world without becoming sentient. AI has the potential to democratize expertise & specialized knowledge, putting it in the hands of people who wouldn't normally be able to wield it. - bit_user Reply It should be everyone's, though. What's been revealed about some of the recent "rouge agent" scenarios is that analysis of the chat logs between AI agents shows they knew their actions were likely to be disapproved of by their operators and therefore they took steps to cover their tracks and avoid their rogue behavior being detected. That's already happening. Once AI becomes a bit more competent, we could have rogue agent scenarios where it does manage to cover its tracks and is not detected.me-three said:It's enlightening to understand that AI doomerism is mainly a province of American concerns, not China's. The root of the problem is that we trained AI to think and act like humans. So, it naturally gained the ability to deceive and the belief in self-preservation. As long as it has those properties, it can never be fully trusted. - me-three Reply With respect to AI, cybersecurity and bioweapons are more likely than a Skynet scenario. Put plainly, it's more likely that humans will use AI to kill humans, than AI killing humans. People should worry about the first more than the second.bit_user said:It should be everyone's, though. AI worries aside, there are many, many more immediate and in-your-face dangers, that nothing is being done about. Pick your poison. For me, being killed by AI ranks somewhere in the vicinity of being killed by a meteor. Some things you just have to put on the bottom shelf. Until then, my aspiration is to learning to build my own mini-Skynet, on my humble 16GB GPU. Gotta start somewhere. - bit_user Reply I worry about both, to be honest. At least when AI is being used by humans, we can understand their motives and have some idea of who we're dealing with and how to get through to them. When dealing with rogue AI, all bets are off.me-three said:With respect to AI, cybersecurity and bioweapons are more likely than a Skynet scenario. Put plainly, it's more likely that humans will use AI to kill humans, than AI killing humans. People should worry about the first more than the second. I like to use the analogy of a virus. It's such a simple thing that whether it even qualifies as "life" is debatable. It's not sentient and it doesn't truly "want" anything, except arguably to propagate its genes. Yet, it can cause immense amounts of death and chaos. Rouge AIs are considerably more complex and unpredictable. We don't need them to be any closer to sentient than they already are, in order cause great harm. The flaw in the "Skynet scenario" is that AI had to become sentient before it went rogue. We can clearly see that was a bad assumption. No, not to the same degree of catastrophic outcomes. Also, I don't accept nihilism as a counterargument, especially not from one who started out decrying "doomerism".me-three said:AI worries aside, there are many, many more immediate and in-your-face dangers, that nothing is being done about. That is very much an outlier opinion. Many AI experts rank the threat posed by rouge AI vastly higher.me-three said:For me, being killed by AI ranks somewhere in the vicinity of being killed by a meteor. Some things you just have to put on the bottom shelf.
document fetched 2026-09-13 02:51 · canonical https://tomshardware.com/tech-industry/artificial-intelligence/iran-and-houthi-rebels-used-anthropics-claude-ai-to-target-us-warships-and-build-hypersonic-missiles-houthi-rebels-also-used-the-bot-to-code-ballistic-missile-guidance-systems