What Happens When AI Becomes a Source of Creative Output?

From videos of what would it be like if the characters of Harry Potter were Balenciaga’s models to winning a photography competition with an AI-generated image, not to mention the sprout of AI-produced cover letters and illustrations for commercial purposes, this technology is literally everywhere. How do we feel about it?

Some years ago, Midjourney and other generative AI programs produced collages and hyperrealistic images far from ever winning prizes and awards. Today, AI-generated visuals are storming the internet with headlines that point at the eery progress of the technology, announcing that what once was a fear of the future, it’s a fear of now: that artificial intelligence is taking over human ability to be creative.

For many years, it has been known that intelligent systems can analyse large amounts of data, facilitating decision-making, optimising processes and reallocating (human) resources in fields like healthcare, finance and transportation. These machines have been programmed to act at the axis of engineering, mathematics, sciences and linguistics in order to identify patterns and solve complex problems based on automation. But what happens when AI becomes, too, a source of creative output?

While many assure that humans won’t get replaced by AI, many cases point to the opposite. For instance, a couple of months ago, illustrator Zhang Wei got commissioned to draw 65 sketches for a novel, a job he’d been carrying out for 8 years. The company was pleased with the first draft and paid right away for it. However, it turned out that that was the first and the last sketch done by Zhang as the company decided to spare money and time by subscribing to an AI program, which they fed with Zhang’s drawing to generate the following sketches. Not only did they easily replace Zhang’s services but the AI program delivered a fairly good job, according to Zhang.

What Happens When AI Becomes a Source of Creative Output?
“Harry Potter by Balenciaga” by Photographer Alexander Niklass with ChatGTP and Midjourney

Cases like this are entering the subconscious of the Internet and creators are panicking over a technology that has the potential to mass produce visuals almost as good as humans do, in much less amount of time and with fewer resources. For instance, Netflix Japan admitted using AI for the background art of the short film The Dog and the Boy and it has sparked quite some backlash as Japanese animation workers are highly skilled and important to the country’s animation industry as they set the tone of the scenes, yet they’re vulnerable to automation.

And so, the first question that pops into the mind of many is, can AI replace creators? Yes and no. To some extent. Probably. And not really.

AI systems function based on the information they are given, and they therefore rely on people’s previously produced knowledge and content. In other words, if it was not because of the artistic output that feeds the machines, AI programs wouldn’t be able to generate artworks or illustrations. Take as an example Matisse and the large amount related to his artworks available to the public. Anyone wanting to generate an image resembling Matisse’s style can do so in the blink of an eye. That’s how the fashion video of Harry Potter and Balenciaga turned out so well as there are plenty of references for AI programs to generate such a refined piece.

What Happens When AI Becomes a Source of Creative Output?
AI generated art of Henri Matisse using wombo.art with the prompt “Henri Matisse miss Matisse Madras 1907” – via Pinterest

But that also means that machines aren’t creative on their own and cannot produce anything from scratch. Their function is not to be creative or more human but to optimise processes in order to catalyse human creativity. But as they’re becoming more sophisticated and threaten to take away the job of many creators, a more important question rises, who owns the rights of AI generated content, which is based on previous content? Many advocate that copyrights should be given to artists whose work seems to be the most influential. But how does that translate financially? At the end of the day, no matter how financial reparations are settled, those using generative systems will most likely reap the benefits — by sparing time and money and getting the content they need for their business.

Although the discussion is centred mainly around the way AI threatens skilled people, for many the conversation should be around the way AI offers tools to democratise and escalate creativity. By mimicking abilities that many have no access to, AI is lowering the high bar by offering the tools to bootstrap the learning of new skills. While these are two opposite takes on the same conversation, both equally important to reflect upon, what’s paramount to acknowledge is that AI is here to stay and it is up to us to establish ways in which we can cope with the technology available, by learning new skills and establishing new categories for AI-collaborations. Competition will increase considerably and many jobs will become obsolete but creativity won’t become AI’s full-time job.

*Header image: ‘The Dog and the Boy’ by Netflix Japan