SynthID watermark is hard to break, but it doesn't solve AI disinformation

SynthID watermark is hard to break, but it doesn't solve AI disinformation(arstechnica.com)

4 points by saikatsg 24 days ago | 1 comment

Swap out "AI" -> "Photoshop" and swap "disinformation" -> "hoaxes"

These problems have been around forever. Instead of trying to detect AI, I really think it's more effective to try to use these tools to make a convincing argument why they are false.

umariomaker 24 days ago |

Swap out "AI" -> "Photoshop" and swap "disinformation" -> "hoaxes"

These problems have been around forever. Instead of trying to detect AI, I really think it's more effective to try to use these tools to make a convincing argument why they are false.