AI Is Starting to Predict Reality… Not Just Words

This week on Slow Takes we break down the rise of AI “world models,” the debate around AI consciousness, and why tools like ChatGPT may still be unreliable in areas like therapy, education, and security. 📰 Substack: https://exploringchatgpt.substack.com 🏆 Guest: https://open.substack.com/pub/theslowai?r=1yyjod&utm_medium=ios Let’s work together! Brand, sponsorship & business inquiries: exploringchatgpt@gmail.com 00:00 Introduction 01:20 Yann LeCun’s $1B Bet on “World Building” 04:10 Why World Models Are Different From LLMs 07:05 Could World Models Lead Toward Consciousness? 12:25 Should We Take AI Consciousness Seriously? 18:35 The Ethics of Conscious AI 24:10 AI Therapy and the Problem of “Deceptive Empathy” 30:40 Should AI Ever Be Used for Mental Health Support? 36:55 Universities Still Don’t Have Good AI Policy 43:20 Why AI Detectors Still Don’t Work 49:05 OpenAI’s Promptfoo Acquisition and AI Security Risks 55:30 Anthropic, OpenAI, and AI Agent Safety 01:00:40 Substack Recommendations 01:03:10 Final Thoughts