Generalist models & Iceman's voice
In this “fully connected” episode of the podcast, we catch up on some recent developments in the AI world, including a new model from DeepMind called Gato. This generalist model can play video games, caption images, respond to chat messages, control robot arms, and much more. We also discuss the use of AI in the entertainment industry (e.g., in new Top Gun movie).
Sign in or Join to comment or subscribe
Thanks for another great episode! Your comments at the end about the legality of using AI to invent new dialog from Val Kilmer’s character were disturbing (but appropriate). As a music composer, I am very sensitive to the legal aspects of copyright violations, and how “close” a piece of music can be to another before it’s considered violating copyright. This is similar, but at a whole level deeper, since it involves directly cloning a person’s likeness. I know there are several applications that will seem useful with this application of AI, but in general I hope the laws get put into place that are very strict against this practice. Just as we’re seeing in the record and composing industries, and the trend for composers and musicians to have to sign away their creative ownership and royalties from their work - so I fear that will follow in the movie industry: sign away your right to your own image and likeness so they can avoid paying you in the future.
Long post, but definitely a huge topic that merits discussion by the country’s lawmakers.
Thanks for the great comment @pwiksell! This is a great summary of the issue, and I definitely think that this will be a very difficult thing to wrestle with as we look towards the future. You could think about this sort of technology being very beneficial (e.g., allowing someone that lost their voice due to throat cancer to continue to “talk” with their explicit permission), but what happens when a voice is used without permission (maybe even with malicious intent). What specific laws apply here? Would love to learn more about this.