This is very good to hear. I've been thinking for awhile now that AI tools need some sort of cryptographic signature in their content to identify the source. This will help protect people in the event of fake images or videos that could be compromising to them, particularly as these tools become more powerful. I guess the next question is how will attackers attempt to fake the watermark.

I just wish that something similar could be done with text content.