I wrote a thing about how we might watermark AI-generated text that works across apps, platforms and devices. It’s a simple/elegant hack for a real problem, and compared to the alternatives 𝙨𝙤 𝙙𝙖𝙢𝙣𝙚𝙙 𝙚𝙖𝙨𝙮.
https://www.wired.com/story/to-watermark-ai-it-needs-its-own-alphabet/
The idea is: Use Unicode to mark text that AI has written. Think “track changes, but for AI.” Not an enforcement mechanism, and voluntary.
Lots of work to do but an elegant hack may be hiding in plain sight.