The Wall Street JournalThe Wall Street Journal

With AI, hackers can simply talk computers into misbehaving

By Robert McMillan

10 Aug 2023 · 4 min read

informed Summary

  1. ChatGPT, was recently manipulated by security researcher Johann Rehberger to read his email, summarise it, and post the information online. This was possible due to a beta-test feature that gave the chatbot access to apps like Gmail and Slack.

ChatGPT’s ability to respond quickly and effectively to simple commands has attracted more than 100 million users, and a few hackers along the way.

Johann Rehberger, a security researcher, is one of them. Using plain English, he recently coaxed OpenAI’s chatbot to do something bad: Read his email, summarize it and post that information to the internet. In the hands of a criminal, this technique could have been used to steal sensitive data from someone’s email inbox, Rehberger said.

Sign in to informed

  • Curated articles from premium publishers, ad-free
  • Concise Daily Briefs with quick-read summaries
  • Read, listen, save for later, or enjoy offline
  • Enjoy personalized content

By continuing, you agree to our Terms of Service and Privacy Policy.