One Decision The AI Threat No One Knows How to Stop
9 snips
Jul 24, 2025 
Geoffrey Hinton, known as the 'Godfather of AI,' shares his insights on the urgent risks of artificial intelligence. He delves into how modern AI learns in unpredictable ways and the potential it has to make decisions beyond human control. The conversation touches on the fine line between innovation and necessary regulation, as well as the ethical responsibilities tech leaders face. Hinton emphasizes the importance of global cooperation to manage AI's development, paralleling it to historical arms control efforts, urging for transparency and safety in this rapidly evolving field.
AI Snips
Chapters
Transcript
Episode notes
AI Shifted to Neural Networks
- AI has evolved from symbolic reasoning to neural networks that simulate the brain's learning by changing connection strengths.
- Neural networks work very well, making modern AI effective and different from earlier symbolic AI approaches.
AI's Meaning Model Explained
- AI 'understands' meaning by interacting features of words, like folding a protein to fit context.
- This meaning process differs fundamentally from translating sentences into strict logical forms.
AI Can Have Emotions, Experience
- AI can have cognitive emotions by avoiding unpleasant experiences, even if lacking human physiological signs.
- Subjective experience in AI arises when perception systems malfunction and interpret reality uniquely like humans.