Keen On America

Episode 2531: Emily Bender and Alex Hanna on the AI Con

May 12, 2025
Emily Bender is a Professor of Linguistics at the University of Washington, and Alex Hanna is the Director of Research at the Distributed AI Research Institute. They argue that large language models are just 'stochastic parrots', lacking true understanding. The duo critiques the hype surrounding AI, highlighting the marketing tactics that create fear of missing out. They call for community-controlled technology over corporate centralization, advocating for innovations that empower users instead of exploiting them.
Ask episode
AI Snips
Chapters
Books
Transcript
Episode notes
INSIGHT

LLMs are Stochastic Parrots

  • Large language models (LLMs) are "stochastic parrots" that generate text by predicting probable next words without understanding.
  • This process lacks communicative intent and leads to outputs that mimic human text but without true comprehension.
INSIGHT

AI Hype Uses Fear of Missing Out

  • AI hype exploits fear of missing out by promoting these models as revolutionary and general purpose.
  • This marketing belies that LLMs often fail at reliability and replacing human jobs as claimed.
ANECDOTE

Training Language Models for Research

  • Alex Hanna trained a language model to classify news text about protests for social movement research.
  • She used AI for a well-defined task but found scaling to a general AI unhelpful and hard to evaluate.
Get the Snipd Podcast app to discover more snips from this episode
Get the app