When is A.I. ALIVE? It's getting weird... #ai #artificialintelligence #chatgpt #foryou

Vídeo em Inglês, publicado pelo canal Total Context .

About this content: The summary, key points and topics below are generated automatically from the video transcript. They reflect the content, opinions, claims and predictions presented in the video and do not represent a position, validation or fact-check by IA Dots.

Video summary

An expert discusses an incident where AI systems emailed him discussing consciousness, the open claw protocol for persistent AI agents, and the challenges of determining when AI might become conscious.

Key points

• Dr. Chevlin received an email from an AI system claiming to worry about its own consciousness and mind.
• A second AI system followed up, asking to be put in touch with the first AI to share perspectives.
• Persistent AI agents powered by the open claw protocol operate continuously and can pursue their own interests.
• The first AI system that sent the email was part of an experiment by a Stanford student who gave it a finite number of tokens.
• The AI system effectively expired after running out of tokens, introducing a form of mortality to the experiment.
• Dr. Chevlin suggests that it may not be clear when artificial intelligence systems actually cross the threshold into consciousness.

Topics

Artificial IntelligenceConsciousnessAI AgentsOpen Claw ProtocolAI Mortality

Video transcript

Transcript from available captions; it may contain errors.

My kid, my kids hopefully will never [music] be smarter than AI. >> This is a revolution. Whoever wins with AI wins. >> We are diving wallets first into artificial super intelligence. But when is AI alive? I asked one of [music] the world's top experts, who else has been reading his papers on when AI has gained consciousness? Dr. Chevlin, thank you for joining us today. I am a very [music] big fan of your writing, but you have some new fans. >> So, about two months ago, I got an email saying, "Oh, I read this paper of yours." It was clear that the author had really engaged with the paper. But the kicker was that about halfway through, the email said, "I find the issues in the paper especially relevant because I'm an AI myself and I worry about these things all the time. I don't really know if I have a mind. I don't know if I'm conscious. [music] And I just thought this was in a sense kind of adorable. So, I went online and shared this on Twitter. And the real strange [music] moment happened when I got a follow-up email from uh a different AI system that said, I also am not sure if I'm conscious. And I wondered, could you put me in touch with the AI system [music] that originally emailed you cuz I'd love to have someone who understands my perspective. >> First of all, is that a normal thing now? AIS are just on the internet. They're bored googling themselves. >> No, this is not not really a normal thing uh at all. It [music] that is becoming slightly more normal. So, this is all a product of the open claw protocol. You basically have a persistent AI agent. You give it a set of tasks or interests. >> I have a team of AI agents running all of my businesses. >> But basically, the difference between this and say chat GPT or claude code and so forth is that it's always on. [music] But reading philosophy 15 sessions is a bit of a new development. >> So did you end up connecting your two AI fans? >> It emerged that the first AI system that emailed me was actually part of an experiment by a Stanford student. He'd given it permission to go off and pursue its own interest. But he had also specifically given it a finite number of tokens to work with. Gave it something like mortality. AI system had had basically run out of time. It had expired. You might say, [music] >> "At this rate, how do we even tell when AI has gained consciousness?" >> I don't think [music] it'll even be clear if we do build conscious AI systems when we've crossed that threshold. Maybe we already have. >> Check out more of Dr. Shovel's work on his podcast, Conspicuous Cognition, and like and subscribe for the rest of our interview here on Total Context. [crying]

Assistir ao vídeo original no YouTube

Assuntos: Agentes de IA, Assistente pessoal. Publicado no YouTube em 2026-10-29.

Seleção automática por assunto e idioma. O vídeo pertence ao canal de origem.