The draft EU regulations said that labelling deep fakes in a programme’s end credits would not be enough, and that viewers of AI generated deep fake content would have to be warned at the beginning and “also at later stages or persistently.” So when the final regulation that appeared today, only said that warnings only needs to be appear “where possible” and “as appropriate”, there was some surprise. And the new EU regulation also included numerous exceptions, such as:

1. Where AI systems perform an assistive function for standard editing
2. Where it is part of an evidently artistic, creative, satirical, fictional or analogous work
3. Where the AI-generated content has undergone a process of human review or editorial control

It was also unclear how it would all be enforced, with many countries yet to specify which national bodies would be in charge of implementation

Tech TV presenters Pete Warren and Bill Mew were joined by legal expert Mark Deem, to discuss the implications