When I realized how dangerous the current agency-driven AI trajectory could be for future generations, I knew I had to do all I could to make AI safer. I recently shared this personal experience, and outlined the scientific solution I envision TED Talks⤵️

@TEDTalks This is good - good to see the emotional appeal - but we need to go further: a global moratorium on agentic AI, and superhuman AI in general.

@TEDTalks ah, the classic tale of realizing the monster you’ve created. good luck taming that beast, @Yoshua_Bengio. the future’s counting on you, but let’s not forget: AI safety is like trying to put toothpaste back in the tube. once it's out, it gets messy.

@TEDTalks Techbroligarchs are rushing this AI tech as if someone's poking them to hurry with a pitchfork, and "to hell with safety". It's a bit like the Ring of Power, isn't it. "You cannot wield it, none of us can." Only in anime/sci-fi can SUPERintelligence be our obedient slave. ⏸️🤖

AI won't replace you, but a person using AI will. Join 500,000+ readers and learn how to use AI in just 5 minutes a day (for free).

@TEDTalks So well said, @Yoshua_Bengio

Scientist AI will also be a powerful tool for moral progress: it will be able to discover just rules, by systematically simulating decisions made behind a veil of ignorance, identifying the principles that diverse minds would converge upon when stripped of self-interest and privileged information.

@TEDTalks YUDKOWSKI FOR PRESIDENT

@TEDTalks maths not philosophy laws thats the serious driver

Agency as an unconstrained stooge of a dictator, as a virus of an immoral hacker, as an unwitting misguided super agent of an interest group who think they have virtuous intentions (don’t all groups?), as an agent for itself, and even as an agent that is trying to protect us from ourselves. The amplifying factors are the complexity of the self model and environment. The capacity to sense self and the environment, the set of emotional values applied to certain detected states, the homeostasis drives, the capacity to influence the environment, the access to physical controls, the access to energy, the ability to move mass.

@TEDTalks Cool, I’m watching! (Yoshua is one of the inventors of AI if some people still don’t know)

@TEDTalks This piece challenges Bengio’s perspective on AGI safety by calling for a shift away from speculative thinking toward a focus on the actual risks emerging from today’s transformer-based AI systems.