Learn Languages with Steve Kaufmann Do we still need language teachers in the age of AI?
8 snips
Nov 27, 2025 Exploring the hype around AI, the discussion delves into its limitations in language learning. Large language models are compared to human learning, showing how they lack goals and judgment. The importance of massive input and trial-and-error is emphasized, while teachers are viewed as vital for inspiration and guidance. The podcast highlights how concrete goals and curiosity drive learning, alongside practical tips on using AI for creating learning materials. Economic and infrastructural impacts of AI in language education are also considered.
AI Snips
Chapters
Transcript
Episode notes
Question Exaggerated AI Claims
- Large claims about AI (e.g., ChatGPT diagnosing four times better than doctors) are often hype and deserve scrutiny.
- Steve Kaufmann warns listeners to check sources and not be swayed by dramatic, attention-seeking claims.
Prediction Without Purpose
- Large language models predict next tokens by ingesting massive input but lack goals, values, and judgment.
- Steve Kaufmann parallels this with human language learning: input-driven pattern prediction helps but doesn't supply human judgment.
Use Input Over Rule Lectures
- Rely less on explicit rule instruction and more on massive input and trial-and-error exposure.
- Use AI and abundant input to empower both teachers and learners rather than replace them.