Emily M. Bender
Linguist; co-author of "On the Dangers of Stochastic Parrots"
Professor of linguistics at the University of Washington and co-author of the Stochastic Parrots paper and the book The AI Con. She argues language models should be described as text generators rather than as understanding systems.
Read them in their own words
Published elsewhere, not by me — so you get real context beyond the name and the title.
What Emily Bender Really Meant by "Stochastic Parrots"
She corrects what people think her famous paper said — worth reading before you quote it.
Key arguments & positions
- Argues large language models are text generators trained on form, not meaning, and that describing them as understanding misleads users and policymakers.
- Co-authored the Stochastic Parrots paper on the environmental, bias and documentation costs of ever-larger training sets.
- Argues "AI" is often a marketing term that hides who is accountable for a decision.
Accomplishments
- Professor of linguistics and director of the computational linguistics program at the University of Washington.
- Co-author of "On the Dangers of Stochastic Parrots" (2021), the paper whose fallout led to Timnit Gebru's departure from Google.
- Co-author of The AI Con (2025) and co-host of the Mystery AI Hype Theater 3000 podcast.
Papers & key writings
Links
In the library
Nothing of theirs is filed in the hub yet. Browse the full library.
Recommended next
Hand-picked from the hub based on what Emily M. Bender covers.
CausalMix: Data Mixture as Causal Inference for Language Model Training
arXiv preprint (July 2026) proposing a causal-inference framing for choosing data mixtures in language-model training.
Why this: Also about language models
AI Is Entering Its 'Slurp Juice' Era
Gizmodo article by Webb Wright (October 7, 2026) comparing the current AI boom with the 2021–2022 NFT and crypto hype cycle, including celebrity promotion and companies appointing metaverse executives.
Why this: Also about AI hype
AI Tells: How AI Writing Differs from Human Writing (Graphite Research)
Graphite research comparing 10,000 human articles with 90,000 AI-generated articles from nine models on the same topics. Measures usage rates of well-known AI tells, corrective framing and intensifiers across GPT, Claude and Gemini model versions, relative to human writing. Published September 2026.
Why this: Also about language models
How to talk about "AI" without adding to the anthropomorphization
Emily M. Bender and Alex Hanna's Mystery AI Hype Theater 3000 newsletter on word choices that stop us describing software as if it thinks, feels or understands.
Why this: Also about AI hype