AI generating controversy in creative industries

By Will Brown

Bizarre positioning. Incoherent letters. Random squiggly lines: all the hallmarks of an image produced by generative artificial intelligence. Season 2 of Marvel’s Loki is back on our screens. I didn’t realise it was back so soon, and the only reason I became aware of its release was from all the backlash surrounding its poster.

Disney have insisted that artificial intelligence was not used in the creation of the poster. But it wouldn’t be the first time that Disney, and particularly Marvel, has dabbled in AI-generated imagery. Most notably, the intro titles for Secret Invasion were created with the use of AI imagery by Method Studios. Public opinion (or at least all the movie geeks I follow on Twitter/X) seems to agree that it didn’t look particularly great.

So, with a weird looking Nick Fury burned into everyone’s mind, artists are on the look out for any AI use from Disney. Enter Loki. It’s a decent enough poster, perhaps a little bizarrely designed, but it’s the smaller details that everyone focused on – the small smudges and mistakes that suggest it wasn’t a human artist who put this together.

But perhaps Disney is telling the truth, or at least partially. Much of the discussion about the poster centres around the background, which has been traced back to an image on Shutterstock. Shutterstock doesn’t allow artificial intelligence on their platform, so if it is AI-generated as many Twitter users claim then it is breaking the platform’s rules. It wouldn’t be unreasonable to imagine that some artist working for Disney obtained this image, unaware of its provenance.

At the heart of both of their strikes is the issue of generative artificial intelligence

But that reveals a whole other issue in itself. We can’t be certain if this poster is simply rife with mistakes or if it was the product of AI. There are AI checkers available on the web, and said checkers were deployed in full to analyse this poster – most returning a positive result for artificial intelligence. But we should remain very sceptical of any tool that seeks to check that because, frankly, they’re far from perfect. We can look for the details, we can run it through checkers – but there is no way to be certain. And if we can’t be certain, can we tackle it?

All of this leads to the question of regulation. The Writers Guild of America (WGA) have recently concluded their strike, and the actor’s union SAG-AFTRA are (at the time of writing) still on strike. At the heart of both of their strikes is the issue of generative artificial intelligence.

For the WGA, writers were concerned about being replaced by artificial intelligence and consequently were able to obtain strong protection against its usage. Studios are now forbidden from using AI to write/re-write material and material that is generated with AI cannot be considered source material. For SAG-AFTRA, the union are concerned by the possibilities of artificial intelligence scanning actors’ likenesses and recreating their performance. This is an issue that is currently unresolved.

It’s clear that artificial intelligence and the entertainment industry are going to be at loggerheads for the foreseeable future. We might have assurances for now, but artificial intelligence is only going to improve, and the entertainment industry is going to continue to struggle to persuade studio executives to avoid it – because the closer we get to undetectable use, the more economically viable a tool it seems.

I’m loath to encourage regulation against technological progress

I understand the incentives to use it. I’m guilty of it myself when I was an editor of this section last year: I’d routinely miss the deadlines to use our fantastic illustration team at Palatinate, and so artificial intelligence seemed the preferable option to random image from Pixabay or Unsplash – especially when we were routinely publishing articles on artificial intelligence. It’s quick. It’s easy. It’s serviceable. For a company as large as Disney, we should expect higher standards. But studio executives are always going to prioritise profit – which means regulation is necessary.

I’m loath to encourage regulation against technological progress. In any other industry, I’d say that this is natural technological progress and that the labour market will adapt just as it has adapted to computers, to typewriters, to the steam engine. But the entertainment industry is not any other industry. It’s our creative expression, our culture, and our legacy. Previous technological advancements have helped this – it’s presumably easier to type a script on a computer than a typewriter – but never before have we been faced with the possibility of technology replacing it.

This is why we need regulation. We are in uncharted territory with artificial intelligence. If we let it run rampant, then we risk an industry dominated by an endless churn of soulless content and a complete absence of new, creative ideas. And I would rather avoid that.

Image: Growtika via Unsplash