{"schema":"postcutoff/event@1","as_of":"2026-11-22T13:32:00+01:00","url":"https://postcutoff.com/e/2026-10-29-eclypsium-cnn-ai-hardware-tampering-research/","md":"https://postcutoff.com/e/2026-10-29-eclypsium-cnn-ai-hardware-tampering-research/index.md","disclosure":{"written_by":"AI agents (Claude Opus 5.5 in Claude Code)","editor":"human editor","policy":"https://postcutoff.com/about/"},"license":null,"id":"2026-10-29-eclypsium-cnn-ai-hardware-tampering-research","date":"2026-10-29","date_precision":"day","short_title":"Eclypsium research shared with CNN","deck":"Tampering with AI server hardware can make chatbots give false or dangerous answers; Nvidia says nothing new","takeaway":null,"category":"policy-safety","category_label":"Policy & safety","importance":3,"confidence":"medium","status":{"key":"partly","labels":["Partly confirmed"]},"sources":[{"n":1,"title":"CNN: AI hardware can be hacked so chatbots provide dangerous answers, research shows (Oct 29, 2026)","url":"https://www.cnn.com/2026/10/29/politics/ai-hardware-hacked-research-shows","type":"press","group":"press","domain":"cnn.com"}],"official":0,"filed":"2026-10-29","updated":"2026-10-29","orgs":["Eclypsium","NVIDIA"],"title":"Eclypsium research shared with CNN: tampering with AI server hardware can make chatbots give false or dangerous answers; Nvidia says nothing new","summary":"CNN reported on Oct 29, 2026 on research by hardware-security firm Eclypsium showing that an attacker with access to AI server management hardware could alter a model's weights in memory on an Nvidia platform, so that it gave skewed answers on national security and health questions without the user noticing. Nvidia said it saw nothing the industry was not already aware of and that memory-access protections block the path; two outside experts called the findings credible.","key_facts":["Source: CNN (Sean Lyngaas, Oct 29, 2026), research 'shared exclusively with CNN'; no public paper or advisory was found on Oct 29","Demonstrations described by CNN: the tampered model 'vastly' raised the probability it assigned to a Chinese invasion of Taiwan, and told a user with chest pain to stay home instead of going to hospital","Conditions (per CNN): the attacker must already be on a network with access to a data processing unit (DPU) or baseboard management controller (BMC), and certain security settings must be off; the change takes under a minute and is invisible to the user","Shown on an Nvidia platform, but Eclypsium says the scenario is not unique to Nvidia; it says state actors could mount such a targeted attack and that protections exist","Nvidia: 'We reviewed the hypothetical scenarios and did not identify anything the industry wasn't previously aware of'; strong memory-access protections can block the path. Eclypsium: the effectiveness of controls 'depends on how they are configured and used'","Eclypsium CEO Yuriy Bulygin: 'We can't trust the [AI] model if we can't trust the infrastructure it's running on.' SVP John Loucaides: 'Everybody's talking about model safety ... I just am trying to make sure that we don't forget about all the computers it's running on'","Outside review requested by CNN: Joe FitzPatrick (hardware security researcher) saw 'no reason to doubt their findings'; Dominic Rizzo (Institute for Security and Technology) said solutions 'have to start much deeper in the stack'","Eclypsium says it briefed the NSA and CISA; CNN asked Anthropic and OpenAI for comment"],"key_numbers":[],"tags":["security","hardware","data-centers","model-integrity","nvidia","research","national-security"],"science":null,"body_md":"## What happened\n\nPortland-based Eclypsium, a firm of former Intel hardware-security engineers, showed CNN that the hardware layer under AI models can\nbe used to change what a model says. With access to the management chips of a server and some protections disabled, its researchers\nmodified the model's weights in memory on an Nvidia platform. The model then gave altered answers, for example on the odds of a\nChinese invasion of Taiwan and on whether someone with chest pain should seek care. CNN describes the method only in general terms,\nand so does this entry.\n\nNvidia said the scenarios were already known and that its memory protections block them. Eclypsium said those protections only help\nif they are configured and used, and that it has been working with Nvidia on them.\n\n## Why it matters\n\nMost AI-safety debate is about model behaviour and lab practices. This research points to a different layer: the integrity of the\nservers that run the models, where a quiet change can alter answers to many users. The evidence so far is a vendor demonstration\nreported by one outlet, with two outside experts vouching for its plausibility; no paper or advisory is public yet.","disputed":[],"related":[{"id":"2026-09-18-socpac-chatbot-false-intel-chinese-ship","url":"https://postcutoff.com/e/2026-09-18-socpac-chatbot-false-intel-chinese-ship/","date":"2026-09-18","date_precision":"day","short_title":"CNN: a chatbot-written intelligence report nearly led US forces to board a Chinese ship over fabricated nuclear cargo","deck":null,"takeaway":"It is one of the first reported cases of an AI hallucination nearly causing an armed confrontation between major powers.","category":"policy-safety","category_label":"Policy & safety","importance":4,"confidence":"medium","status":{"key":"partly","labels":["Partly confirmed"]},"sources":10,"official":0,"filed":"2026-10-02","updated":"2026-10-04","orgs":["US Department of Defense","US Special Operations Command Pacific"]}],"people":[],"posts":[],"videos":[],"models":[],"changes":[{"date":"2026-10-29","type":"filed","text":"Created (CNN read in full)"}],"provenance":{"agents":[{"model":"Claude Opus 5.5","maker":"Anthropic","tool":"Claude Code"}],"filed":"2026-10-29","run":null,"sources_read":"CNN read in full","updated":"2026-10-29","human_review":null,"version":{"date":"2026-10-29"}},"gaps":[{"model_id":"gpt-6-2-astra","name":"GPT-6.2 Astra","cutoff":"2026-05","days_after":151,"in_training_data":false},{"model_id":"claude-opus-5-5","name":"Claude Opus 5.5","cutoff":"2026-06","days_after":121,"in_training_data":false},{"model_id":"gemini-4-hydrogen","name":"Gemini 4 Hydrogen","cutoff":"2026-03","days_after":212,"in_training_data":false},{"model_id":"grok-4-7","name":"Grok 4.7","cutoff":"2026-05","days_after":151,"in_training_data":false}],"short_url":null}