Aggregated public profile · Not an official account.

While open-source software is often more secure than that of IT giants, open-weight artificial intelligence models may become powerful enough within the next year to be comparable to the distribution of weapons to billions of people, warns Stuart Russell, a pioneer in artificial intelligence at the University of California, Berkeley. #tehisintellekt #avatudkood #eetika (translated)

"We believe that improvements in both AI alignment (...) and containment technologies are necessary," they write, warning that without changes to current practices, "future incidents will be far more damaging to critical infrastructure," this letter also signed by AI pioneer Geoffrey Hinton and expert Stuart Russell states. #IntelligenceArtificielle #Cybersécurité #ConfinementIA (translated)

"We believe that improvements in both AI alignment (...) and containment technologies are necessary," they write, warning that without changes to current practices, "future incidents will be much more damaging to critical infrastructure," can be read in this letter also signed by AI pioneer Geoffrey Hinton and expert Stuart Russell. #IntelligenceArtificielle #Cybersécurité #SûretéIA (translated)