Hello everyone!
In 2021 the European Commission has proposed the EU AI act, a first attempt of regulating AI at European Union level and about 2 years later it seems like the proposal will soon be adopted by the EU, that is why I thought it would be the perfect time to reflect a bit on this Act and specifically its provision related to deep fakes, this is, again, the first time the EU is attempting to regulate on deep fakes.
Some criticism has already been written on the proposed regulation of deep fakes, which include the fact that there not a clear definition of what is a deep fake; that the use of deepfakes is allowed by the proposal, but deepfake technologies are considered a minimal risk AI system meaning that only a transparency obligation applies, this means that it needs to be disclosed that deep fake technology was used, however most deepfake creators are anonymous as the deepfakes are created with a malicious intent, so one can wonder if a transparency obligation is enough for all types of deep fakes!
I will leave you two articles I have found interesting that reflect on the challenges of regulating deep fakes in the EU AI act!