We have SSL for website authenticity, DKIM/BIMI for email authenticity, and we're getting closer to C2PA provenance for content authenticity.

'watermarking AI will fail' is correct, but because it's trying to prove a negative (stop bad actors) instead of amplifying the good (those who just want to verify they are who they say they are)

Building Chamber. Prior: VP at Adobe, Head of Behance, co-founder of the Content Authenticity Initiative, early days at TED. https://withchamber.com/

'watermarking AI will fail' is correct, but because it's trying to prove a negative (stop bad actors) instead of amplifying the good (those who just want to verify they are who they say they are) https://twitter.com/emollick/status/1723842108484645046

In some areas it is still theoretical vs real. On the political front, it isn't a real problem IMO - it's all anticipatory. On the war front, it is - a clear infowar happening trying to build support or sympathy for Hamas. Deepfake pornography is an entirely different problem with real ethical quandaries