EU AI act, Watermarks and labeling Gen-AI

Matt Hall will reflect on the AI act and labeling Gen-AI work. Article 50 of the EU AI Act, which came into effect on 2 August 2026, introduced transparency requirements for providers and deployers of all AI systems, not only those in the high-risk category. Depending on the situation, requirements may include visible human-readable labels and invisible machine-readable watermarks.

In this talk, Matt will show various approaches people are using, including IBM's AI Attribution, C2PA Content Credentials, and Google's SynthID-text, and discuss some of the ethics of labeling and watermarking.

(Spots are limited and people end up on a waitlist, so register early!)

We meet at Visere office in [Chr. Michelsens Gate 1](https://www.google.com/maps/place/Visere+Bergen/@60.3937601,5.3228291,19.45z/data=!4m6!3m5!1s0x463cff9d94c38b0d:0xd0dbf334ab2db287!8m2!3d60.3937768!4d5.3232435!16s%2Fg%2F11w7_p3f6b?entry=ttu&g_ep=EgoyMDI1MTAxMy4wIKXMDSoASAFQAw%3D%3D). It says T.W.R. on the doors, and there is a number you should call. If you are struggling to find it, send me a message here on the meetup (Dalibor Jovic) with your phone number, and I will help you up.