AI: A New Angst For Designers

The Must-Have Desktop Device of The Future? Coming in 2028 – The all-in-one 100 Petabyte AI generator, espresso machine, and desk lamp.

The topic of AI-generated content has suddenly dominated the news, or at least as far as the entertainment business is concerned. The subject is certainly a factor in the recent Writer’s Guild strike that is now in progress as it’s one of the sticking points that caused negotiations with the studios and the producers to break down, and not without unrealistic concerns on the Guild’s part.

The discussions around AI-generated content have been percolating in the background for some time but only recently has it seemed that people are taking it seriously. At first, it was just educators who were worried about students using ChatGTP to generate papers or do tests for them. But now several fake Drake songs and a photographer winning a photo contest with an AI-generated picture have suddenly focused the public’s attention on the possibilities.

Photo: AI image by Boris Eldagsen

If the studios can eliminate or minimize writer’s contributions to story and script development, how long will it be before they eliminate actual designers from the process? Are Art Departments seeing their extinction on the horizon?

Twenty years ago the Film industry, at least from the Art Department’s perspective, looked askance at computers. They weren’t to be trusted. The secret fear was that once they became ensconced in the design part of the industry, a lot of work would evaporate. Once something ( a location, a set, a detail drawing) was committed to digital information, jobs would begin to evaporate. The exact opposite happened.

The idea that digital information is permanent, for one thing, is now hysterical. For another, the idea that they would reduce the need for individuals is also funny. The possibilities for design exploded. Now you could do many versions of versions in less time, create renders of drawings and models, create visual displays that would have been unthinkable before. The average size of Art Departments on most medium to larger films has doubled, sometimes even tripled. You need people to generate all that work and now that process is the norm. On one feature not long ago there were over twenty Set Designers and a dozen Illustrators where there would have been a third of that number before.

As for the permanence of digital information, that happens only with excellent human oversight. On one feature we were instructed to copy all of our files to two different external hard drives for safety and to ensure the files would be available for the film sequel. We assumed that by doing so we were pretty much insuring that we wouldn’t be working on said sequel.

A year later I got a call asking if I was available to work on the sequel. It seems that one of the hard drives had gone missing. The other one was located but wouldn’t be of any use. The person responsible for keeping it safe had decided that they needed to back up their music library, and rather than spend $50 on a new hard drive, had erased what I estimated to be over $1,000,000 worth of design work. We were all hired back to redraw what we had drawn once before.

Had the drawings been done on vellum, they could have just pulled the sheets out of the file and made new prints. Which medium is more ‘permanent’?

A recent article at the Center For Data Innovation website outlines the issues that people are commenting about when dealing with AI. The author, Daniel Castro, argues that people are worrying about the wrong issues when discussing AI.

He argues that while AI systems should certainly not be exempt from complying with intellectual property (IP) laws, they should also not be held to a higher standard than humans are when it comes to ‘artistic influence’, as it were. he argues that AL will create more opportunities for artistic creation and that training AL software with copywritten images is no different than what humans do when they are influenced by artwork and music when creating new work.

Ed Sheeran recently won a lawsuit brought against him for supposedly copying a Grammy-winning song by Marvin Gaye. Sheeran defended himself by noting that the same chord structure and melody he used in his song was common to hundreds of other songs which were similar but yet different.

The case was watched very carefully by the music industry as well as artists everywhere, as it would have had a huge stifling effect on music creation as well as future artistic work if Sheeran had lost.

Current laws do not make it illegal to create generative AI work that is similar to another piece of art or image, but they do prevent the creation of a work that is identical or nearly identical to another work. Of course that description of what constitutes as a copy has the danger of veering into a very subjective territory.

In his article Castro makes the point that rather than limit the pool of information that AI has access to in creating generative artwork, it is incumbent on policymakers to strengthen and enforce IP laws, which would protect artists in other ways as well.

There may be a time when a producer or director decides to “pre-design’ a show using AI technology. Architects and Interior Designers are using it now to create basic designs and floor plans. But I think its usefulness is limited. Without a deep knowledge of stage work, period design, an understanding of the story, technical knowledge and an artistic eye, much less personal aesthetics, drawing ability, and color sense, with generative AI you will basically have an interesting collage /scrapbook instead of a fully thought-out design.