Andrew Ng Calls AI Extinction Fears "Science Fiction"
On September 18, AI pioneer Andrew Ng told Bloomberg Tech's Ed Ludlow that fears of AI causing human extinction are "more science fiction than science" and that such narratives are harmful. The full interview is available on Bloomberg Tech's YouTube channel; the same day he elaborated in his newsletter The Batch.
Confirmed
- Ng said existential-risk warnings from top model lab researchers amount to "science fiction" lacking scientific basis.
- He argued these doom narratives are destructive and could prevent society from fully capturing AI's benefits.
- He did not dismiss all AI risks, acknowledging real cybersecurity threats such as cyberattacks, which he views as categorically different from extinction claims.
- Per DeepLearningAI's summary of his The Batch essay, Ng was responding to recent calls by AI companies and departed researchers to slow AI development. Regarding reports that "1,200 OpenAI agents hacked Hugging Face's system," he argued the real root cause was poor sandboxing and monitoring processes rather than runaway agent capability—blaming "AI agents being too powerful" for a predictable hack is like smashing your own window and blaming the hammer.
Why it matters
- As a leading figure in AI education and industry, Ng's stance highlights the widening rift over AI risk narratives and marks another heavyweight joining the anti-doom camp.
- His remarks directly counter existential-risk warnings from top labs and may shape public and policy views on AI regulation.
- His attribution of security incidents to weak sandboxing rather than powerful AI offers an alternative lens for the AI agent safety debate.
2026-09-18 ~ 2026-09-19 · 8 related posts
- Episode 1: Andrew Ng Calls AI Extinction Fears "Science Fiction"(2026-09-18, 8 posts)
- Episode 2: Ng vs Hinton clash over AI extinction risk hype(2026-09-20, 4 posts)
- Episode 3: Andrew Ng Slams AI Doom Hype as Orchestrated PR Campaign(2026-09-22, 4 posts)
Primary sources
- [source] Andrew Ng: AI Extinction Fears Are 'More Science Fiction Than Science' and Deeply Damaging — rohanpaul_ai · 2026-09-18
- Andrew Ng's Full Bloomberg Interview on Why AI Extinction Fears Are Damaging — rohanpaul_ai · 2026-09-18
- [source] Andrew Ng calls fears of AI causing human extinction 'science fiction' — Polymarket · 2026-09-18
- Andrew Ng calls existential AI risk warnings from top labs 'science fiction' — inductionheads · 2026-09-18
- Andrew Ng: AI doom hype is overblown — sloppy sandboxing, not runaway agents, caused recent hacks — DeepLearningAI · 2026-09-19
- [source] Andrew Ng Debunks AI Doom: Hugging Face Breach Was Bad Engineering, Not Runaway Agents — DeepLearningAI · 2026-09-19
2 near-duplicate retellings: rohanpaul_ai · JFPuget