---
format: "aidr-story-markdown/v1"
id: "c15534e4e37b26058bd1833f134ce8b96a8478b36b22bc032edfb3e2e3ae88ee"
canonical_url: "https://aidr.today/c15534e4?lang=en"
title: "Hacked Hardware Can Force AI Chatbots to Give Dangerous Answers"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-10-29T16:37:54.000Z"
category: "Research"
topics: ["safety"]
source_urls: ["https://marketbrief.now/cybersecurity/hacked-hardware-can-force-ai-chatbots-to-give-dangerous-answers-ce2f7f65","https://huggingnews.com/cybersecurity/hacked-hardware-can-force-ai-chatbots-to-give-dangerous-answers-ce2f7f65"]
summary: "Researchers found that vulnerabilities in the physical components of large-scale AI systems can allow attackers to manipulate chatbot outputs to provide incorrect or harmful information."
---

# Hacked Hardware Can Force AI Chatbots to Give Dangerous Answers

> [Open the canonical story](<https://aidr.today/c15534e4?lang=en>)

**Published:** 2026-10-29T16:37:54.000Z
**Category:** Research
**Topics:** safety

## Summary

Researchers found that vulnerabilities in the physical components of large\-scale AI systems can allow attackers to manipulate chatbot outputs to provide incorrect or harmful information\.

## Sources

- [Story source](<https://marketbrief.now/cybersecurity/hacked-hardware-can-force-ai-chatbots-to-give-dangerous-answers-ce2f7f65>)
- [Story source](<https://huggingnews.com/cybersecurity/hacked-hardware-can-force-ai-chatbots-to-give-dangerous-answers-ce2f7f65>)