Julian Togelius is an Associate Professor of Computer Science at NYU working on artificial intelligence. He uses AI to make games more sophisticated, and he uses video games to develop smarter AI's. Today we discuss the ethics of open AI, why AI Could destroy humanity, and what we can do to stop it. 00:00 Intro 00:40 Building AI to understand the mind 04:54 Deep learning doesn’t understand truth 07:14 Gaming helped Ai development 09:11 Meta’s AI + computation speed v Brain 14:12 Latest AI + diffusion model & control 18:42 Language models + sophisticated guessing game 24:42 Understanding Artificial General Intelligence 34:25 Consciousness 47:32 ChatGPT is LARPing + Where does responsibility lie? 53:36 Evolving + video games + survival 01:01:28: Desire to propagate + immortality 01:11:01 Using Super Mario for AI + Games in tutorial 01:35:42 AI playing chess + game rule sets 01:45:36 Creating AI curiosity, evolution & playing 01:55:58 Deep connection between gaming + AI 01:56:56 Concerns about AI + printing press impact 02:02:27 “WarGames” best film about AI 02:08:47 AI aligning with humans’ best interests 02:14:21 Centralized planning doesn’t work 02:19:06 OpenAI is hypocritical + should dissolve 02:22:15 People will destroy humanity not AI 02:24:09 Positivity impact of AI - education 02:27:04 Understanding truthfulness 02:28:50 Perfectly unique Media experiences 02:30:59 Innovations - videos, mining/oil extraction