Deep Learning models has been the center of attention for quite a few years now; they have been made fun of for being so costly to train and yet being so poor at some tasks, they have caught attention as they seem to be able to do so many thing in a single architecture yet nobody knows how they work, more recently they have been feared for the potential of deep fake but also generative models such as Dall-E or midjourney.
During the past 6 months, the US are facing a massive strike from artists, actors, etc. as big names in the movie making business are thinking about using generative models to replace actors, voice actors, concept artists, etc.
Once again, this is still software so… Have we missed an opportunity to prevent this? Have we missed an opportunity to make new collaborations and apply our techniques to this new problem?
This might be an idea that I could consider submitting to the “New and Controversial Ideas” track for VaMoS 2024