Actually, a few weeks ago I had to attend a seminar/panel for my study abroad on AI as it pertains to the future of the entertainment industry, and it was really interesting and kind of made me challenge some of my beliefs on AI. It was made up of a few television writers and two producers, and they had some wildly different takes.
The writers were very much against it, saying that it has made their jobs much harder and forced them to have to compete with these giant beasts of programs that were trained on so much work and can pump out content incredibly quickly, that it would be a momentous task to work against its services. Additionally, during the writers’ strike, companies quickly turned to these programs as a way to scab the damage, and it worried them to see the repercussions of the future.
On the producers’ side, they were really excited about the future prospects of the industry using AI, but not for the reasons I assumed. They were excited about the non-generative tools of AI. One thing that stood out to me from the conversation was a script-writing program, where it could analyze a script and produce coverage/an outline, so writers could see if their main points were coming across. Included in the program was a table-reading function, so writers could hear a “realistic-sounding” AI cast read out their script to see if the dialogue held up appropriately. That was something that blew my mind and I think it could be pretty useful, and in this way it made me re-think the use cases of non-generative AI.
Where it gets tricky for me, however, is the case of generative AI. I had asked the producers about the issue of credit, and they were bit, like, wishy-washy about it? They said as it stands there’s no place for generative AI in the industry right now since it can’t be copyrighted, but that it’s the same process a human would go through to create work-- first copying the masters to learn the techniques before producing their own works (which, I’ll say, I don’t really agree with. I think there’s a huge difference between a person sharpening their craft and a machine consuming a huge volume of works to put out something similar). They also compared people’s worry about AI to that of the worry of photoshop, when it first becoming relevant. People thought that photoshop meant the end of photographers’ careers, and yet there’s still a need for them and their artistry. It’s just a tool for creators to use.
So, as it stands, I’m still very against generative AI, although even then it’s complicated. For one of my internships, I have to use generative-fill on photoshop sometimes on graphics to extend the sky, or a wall, or the ground for the photos to fit a certain size. To me, that feels pretty “fine” as a use case, but where is the line drawn? Something to think about, and something I’m still grappling with.