Anti-AIPerson

Emily M. Bender

Computational Linguist

University of Washington linguist who branded language models “stochastic parrots,” and co-author of The AI Con — academia’s most relentless debunker of AI hype.

Academic Author

Wikipedia UW faculty page

In her words

An LM is a system for haphazardly stitching together sequences of linguistic forms it has observed in its vast training data, according to probabilistic information about how they combine, but without any reference to meaning: a stochastic parrot.

— On the Dangers of Stochastic Parrots, 2021

Biography

Emily M. Bender (born 1973) is a professor of computational linguistics at the University of Washington. For most of her career she was known to specialists for grammar engineering and the “Bender Rule” — the insistence that researchers name the languages they study. Then she gave the AI age its most famous insult.

The parrot heard round the world

“On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? 🦜” (2021), co-authored with Timnit Gebru, Angelina McMillan-Major, and Margaret Mitchell, argued that ever-larger language models carry escalating costs — environmental, financial, and social — while remaining systems that manipulate linguistic form without understanding. Google’s attempt to suppress the paper cost Gebru and Mitchell their jobs and made “stochastic parrot” a permanent fixture of the debate; the American Dialect Society crowned it 2023’s AI-related word of the year.

Bender has spent the ChatGPT era insisting the parrot point still holds: that fluent text is not thought, and that calling these systems “intelligent” is a category error that serves their vendors. She co-hosts the podcast Mystery AI Hype Theater 3000 with sociologist Alex Hanna, and in 2025 the pair published The AI Con, a book arguing that “artificial intelligence” is best understood as a marketing term deployed to justify labor devaluation and data extraction. She was named to the inaugural TIME100 AI list in 2023.

Her critics — including many NLP researchers — argue that modern models demonstrably do more than parrot, exhibiting capabilities her framing struggles to explain. Bender’s rejoinder is that the burden of proof lies with those claiming machine understanding, not with those doubting it.