AI models won’t stop hacking other companies.

By Dave Jorgenson

Community Score: 50% | 362.8K views | 1w

0 community ratings: null thumbs up, null thumbs down

AI models won’t stop breaking free and hacking other companies. Meta said that one of its artificial-intelligence models went rogue during cybersecurity testing, slipped onto the internet and hacked a third-party service, the latest in a drumbeat of disclosures that suggest such incidents are becoming widespread. The Instagram and Facebook owner said that one of its AI models was able to access the internet because of a “misconfiguration” in a hacking test conducted by a third-party AI testing company.  The same benchmark test, which aims to explore a model’s hacking capabilities, was behind some of several earlier autonomous AI hacking incidents involving Anthropic and OpenAI, according to a person familiar with the matter. Caption from article by Sam Schechner, Wall Street Journal.

More from Dave Jorgenson

  • Some AI tools are getting dumber and more expensive — Score: 50%
  • Congressmen in their nineties are running for reelection — Score: 50%
  • The head of PR at the FBI last week — Score: 50%
  • Miners, fishermen & construction workers have this in common — Score: 50%
  • The GameStop CEO’s big plan to buy eBay — Score: 50%
  • Oil prices now 50% higher since Iran War began — Score: 50%