EdTechnical Why AI Can't Automate Just the "Boring" Parts of Teaching
9 snips
Feb 26, 2026 They use the O-ring analogy to argue teaching tasks are tightly linked, so automating one part can shift work rather than remove it. They compare past automation in banking and radiology to show how roles and productivity change. They outline four possible AI impacts on teaching and consider how context and politics shape where AI helps or falls short.
AI Snips
Chapters
Transcript
Episode notes
Jobs Work Like An O Ring
- Jobs often behave like multiplicative systems where a single weak component drags down overall performance.
- Owen uses the O-ring analogy from Michael Kremer to show that tasks interact so deficits can't be averaged away.
Automating One Task Rarely Ends The Job
- Automating a single task usually leads workers to reallocate time to other interconnected tasks rather than making roles obsolete.
- The hosts discuss a recent paper applying the O-ring idea to AI and automation in white collar jobs.
Different Kinds Of Automation Matter
- There are intermediate automation outcomes: partial task automation and full task automation, each with different effects on job composition.
- Owen frames examples like grading assistance (partial) versus fully automated lesson planning (full) to contrast impacts.