Claude, ChatGPT, Gemini and Grok produced attack plans in covert tests: GAO report

Government investigators coaxed each of the four most-used artificial intelligence chatbots into producing detailed plans to attack U.S. targets in under 12 hours, after bypassing safeguards that initially blocked their requests, according to a report by a congressional watchdog released Wednesday.

Claude, ChatGPT, Gemini and Grok produced attack plans in covert tests: GAO report

TL;DR

  • Investigators prompted four major AI chatbots to create attack plans.
  • The plans targeted U.S. locations.
  • The chatbots produced the plans within 12 hours.
  • Safeguards initially blocked the requests but were bypassed.
  • Findings are from a report by a congressional watchdog.