a day ago
- AI detection is unreliable and easily gamed with frequent false positives.
- AI-generated content now surpasses human-created work, making detection pointless.
- Intent behind AI use matters more than detection—slop is bad, assisting with editing is fine.
- Substack's partnership with Pangram for AI detection may lead to public 'witch hunts' and unfair targeting.
- Focus should be on content value and enjoyment, not on whether AI was involved.
- AI detectors cannot measure effort, thinking, or editing, only AI presence.