Substitute teaching my partner’s history class last week. The lesson plan used a YouTube video that showed ‘WWII era color footage’ of London streets. Hyper realistic, narrated, looked like authentic restoration work. Spotted that the lamppost shadows were inconsistent with the implied time of day. Ran the video through a deepfake detector and it came back 94% AI. The teacher who picked it had no idea, just searched ‘WWII London street footage’ and grabbed a high view-count video. This is going to happen everywhere. How do we equip teachers to verify before lesson planning?
Were dealing with the same problem at university level. Saw a ‘restored medieval manuscript reading’ video used in a lit seminar that turned out to be AI generated, completely fabricated scribe behavior and a non-existent manuscript. Were piloting a verification checklist for teaching materials. Anything ‘restored’ or ‘newly discovered’ goes through a 3-step check before use. Slows lesson prep but the alternative is teaching students that fabricated content is real.
From the editorial side, this is now a documented issue in K-12 textbook publishing. Several major publishers have pulled video resources after discovering AI generation. There needs to be a Library of Congress or equivalent provenance database for historical media that teachers can default to. Right now its just YouTube which is open season.
this is wild and also explains some stuff in my history class that didnt add up. we watched a ‘restored’ video about ancient rome and the architecture wasnt right for the period and the professor was like ‘ai magic’ but yeah it was probably just ai
I think were going to need media literacy modules baked into teacher training programs. The next generation of teachers needs to learn to verify before they teach. Going to bring this up at my district PD next month.