https://n2v.paloaltoonline.com/square/print/2024/01/15/real-or-fake-bill-would-force-tech-companies-to-watermark-ai-generated-content


Town Square

Real or fake? Bill would force tech companies to 'watermark' AI-generated content

Original post made on Jan 16, 2024

As the field of generative artificial intelligence explodes in popularity, ushering in a torrent of fake media, state Sen. Josh Becker is sponsoring a bill that would require AI companies to watermark their content.

Read the full story here Web Link posted Monday, January 15, 2024, 3:18 PM

Comments

Posted by bill1940
a resident of Menlo Park
on Jan 16, 2024 at 10:28 am

bill1940 is a registered user.

Why stop with videos, audios, and images? One can then have an app that can detect the water mark.

There must be methods to also "water mark" text generated by AI like ChatGPT-4.

The "water mark" will not be like what one can do with the above media. In these cases it is easy to hide hidden bit patterns in the data. With text it might be a few randomly placed text patterns that fit perfectly in the context of the text document but have a very low probability of occurring in human generated text documents.


Posted by ReallyLiveHere
a resident of Fairmeadow
on Jan 16, 2024 at 11:15 am

ReallyLiveHere is a registered user.

Going to be very hard to enforce in a world where open-source software can be run on home computers to generate images.