Went to see Oren Etzioni speak in Pittsburgh recently about AI. He talked about the need for an "impregnable off switch". I Got home and Asked Chat GP if humans could do it? Got a long answer but answer boiled down to, "it's complicated but probably not. Sleep well.
Went to see Oren Etzioni speak in Pittsburgh recently about AI. He talked about the need for an "impregnable off switch". I Got home and Asked Chat GP if humans could do it? Got a long answer but answer boiled down to, "it's complicated but probably not. Sleep well.
Is nearly impossible. AI Safety researcher (like those at OpenAI and Deep Mind) focus on designing AI that aligns with human values rather than just relying on a switch.
January 30, 2025 at 3:57 PM
Is nearly impossible. AI Safety researcher (like those at OpenAI and Deep Mind) focus on designing AI that aligns with human values rather than just relying on a switch.
“an impregnable OFF switch" just in case.... So, I asked Chat GTP the following question, "Can humans create an impregnable Off switch for AI". A well-designed Off switch can work in controlled environments (like narrow AI or local systems), but for highly advanced AI, ensuring 100% effectiveness
January 30, 2025 at 3:55 PM
“an impregnable OFF switch" just in case.... So, I asked Chat GTP the following question, "Can humans create an impregnable Off switch for AI". A well-designed Off switch can work in controlled environments (like narrow AI or local systems), but for highly advanced AI, ensuring 100% effectiveness