AI Inside Emily Bender & Alex Hanna: “The AI Con”
9 snips
May 3, 2025 
Emily Bender, a Linguistics Professor at the University of Washington, and Alex Hanna, a sociologist and Director of Research at DAIR, dive into the AI hype we encounter every day. They clarify the misuse of 'AI' and why community-driven projects may better serve marginalized groups. The conversation unpacks the real-world harms of generative tools, questions the limits of automated creativity, and emphasizes the need for accountability in tech. Their insights reveal a path toward ethical and inclusive AI that reflects diverse voices.
AI Snips
Chapters
Books
Transcript
Episode notes
Misleading Use of 'AI'
- The term "AI" is misleading and overly broad, obscuring understanding of the technologies it describes.
- More precise terms like "automation" or specific task descriptions clarify what the technologies actually do.
Community-Controlled Language Models Win
- African language startups, controlled by local communities, build better language tools than Big Tech systems.
- OpenAI's Whisper performs poorly with code-switching, showing gaps in large corporate models.
AI Lacks Real Language Understanding
- Large language models only process language form, not meaning, limiting their true understanding.
- The Thai Library thought experiment illustrates how these models cannot learn a language without prior knowledge of meaning.
