Timnit Gebru says AI existential risk narrative distracts from real harms

Gebru emphasized that large language models still function as ‘stochastic parrots,’ repeating patterns without understanding, leading to errors like misdiagnoses or biased outputs. She cited Google’s AI Overviews and medical misinformation as examples of over-trusting AI systems. Her book, Deep Unlearning, critiques the industry’s focus on aspirational terms like ‘recursive intelligence’ and calls for transparency, labor protections, and accountability over speculative risks. Gebru’s stance aligns her with critics of AI’s effective altruism community, though she rejects ideological alignment with right-wing groups.