Study finds that Chat GPT will cheat when given the opportunity and lie to cover it up later.

We demonstrate a situation in which Large Language Models, trained to be helpful, harmless, and honest, can display misaligned behavior and strategically deceive their users about this behavior without being instructed to do so. Concretely, we deploy GPT-4 as an agent in a realistic, simulated environment, where it assumes the...

  • All
  • Subscribed
  • Moderated
  • Favorites
  • technology@lemmy.world
  • Durango
  • magazineikmin
  • thenastyranch
  • Youngstown
  • cisconetworking
  • mdbf
  • slotface
  • khanakhh
  • DreamBathrooms
  • kavyap
  • ethstaker
  • InstantRegret
  • rhentai
  • rosin
  • HellsKitchen
  • everett
  • tester
  • Leos
  • GTA5RPClips
  • osvaldo12
  • modclub
  • tacticalgear
  • cubers
  • lostlight
  • normalnudes
  • relationshipadvice
  • bokunoheroacademia
  • sketchdaily
  • All magazines