AI (part three)
I'd planned on having a collection of your email today, but I received one this morning from someone who brought up an excellent rabbit hole to explore. What follows is lightly edited for clarity:
I work at one of the big ad agencies and we aren't allowed to use AI for client work unless it's only utilizing a database of assets (jphotos/videos/graphics) that the client actually owns. The concern is if we use AI that pulls from everywhere we haven't licensed the creative it comes from. We open ourselves to being sued if someone can prove their art was used for something that made it to broadcast or print. In our industry we are leaning very conservative for great fear of plagiarism lawsuits.
The creative work coming through our agency has taken a bit of a dive due to MidJourney. The creative teams can use it internally to generate images to go along with the concepts they are coming up with to sell to clients. This speeds up the process, but it's been a detriment to the quality of ideas. Before, creatives would come up with an idea and create an image for the concept in Photoshop (which would take hours). During that time they would think about the idea/concept and refine it - sometimes improving it because they had to live with the idea for a period of time (also sometimes tossing it out because it wasn't good enough). Now they spend 5-10 minutes and move on. While it's increased the volume of ideas, the quality is lessened as is their understanding. When I ask (as a Producer) "How does this work?" they have no idea.
Here's the thing about creativity: it's time-driven. While ideas often happen out of "nowhere," they've been churning in the background of our minds for much longer. Also, so much of creativity is lateral thinking, not vertical. What AI generation programs for art seemingly do is remove the lateral thinking aspect.
Much is lost.
When I write, I often don't come up with the right phrase or idea until I've gone through 5+ drafts (not infrequently, 10+). Not everyone works that way, but I do. Without that time spent, the text lacks dimension, and dimension is what gives it vibrance.
What AI can do is generate an infinite amount of content, but without the reflection and iteration necessary to drive it to a higher level. For now, anyway.
<< Home