FT Tech Tonic Superintelligent AI: The Doomers
39 snips
Nov 14, 2023 
Yoshua Bengio, a pioneer of generative AI, and Eliezer Yudkowsky, a research lead at the Machine Intelligence Research Institute, discuss the existential risk of superintelligent AI. Yann LeCun, head of AI at Meta, disagrees and points out the potential benefits of superintelligent AI. Topics include the dangers of superintelligent machines, aligning AI systems with human values, and the potential and misuse of superintelligent AI.
AI Snips
Chapters
Transcript
Episode notes
Frankenstein as AI Metaphor
- Mary Shelley's novel Frankenstein tells of a scientist whose creation turns against him with disastrous consequences.
- This story is used as a metaphor for fears about building super-intelligent AI that could harm humanity.
Bengio Warns of Rapid AI Progress
- Yoshua Bengio believes super-intelligent AI is coming within a decade based on recent progress trends.
- He warns society is unprepared to handle the resulting disruptions and misuse risks.
ChatGPT Changed Bengio's View
- The breakthrough of generative AI like ChatGPT shifted Bengio's perspective on AI progress.
- He now considers human-level reasoning AI possible within a few years, not decades.