Back to blog index

Creative industries are under assault by AI

Posted by Vlad Calin 3 weeks, 3 days ago

The hype with the recent AI developments is still going somewhat strong. Even though the initial awe went away and people started to see ChatGPT for what it is, a text generation tool that sometimes gives good answers but most of the time it gives the wrong ones, confidently, the disruption is still under way.

The rise of generative AI is something to really think about.

What does it mean for professionals that are earning a living from drawing and producing graphical content when AI comes in and can do a lot more way quicker and for lower costs for the clients.

First of all, let's just say that AI at this stage, got kind of good generating new things based on stuff it saw before as training data. There are some ethical concerns around that, because basically AI is cannibalizing the "hand" that fed it its training data.

Secondly, AI is still kind of bad at thinking. It is good at finding some patterns in the inputs, and putting together some new stuff based on some patterns it identified as being related to the input patterns. That's how I would sum up AI right now in Layman terms. And it excelled at that. It can identify certain words in the input prompts, and put together some other words or pixels in certain configurations so good, that it generated something that can be used right off the bat as creative content.

So, next let's talk about the two areas where AI is currently disrupting the creative process

Graphical content

When it comes to graphical content, AI is going strong. Illustrations, vectors, designs, drawings, paintings, AI can do that at an acceptable level. And while dall-e, the OpenAI's model is somewhat mediocre because it was a general purpose experiment, there are companies out there that take this general unspecialized model and specialize it on certain niches.

For high quality illustrations, you got MidJourney that somehow managed to get their model to give high quality images from the simplest prompt. The quality of the generated content is astonishing, and it can even generate website or app designs to use as starting points for something real. The biggest struggle AI has at this point in this area is putting text in the images. Even if you ask it to put a specific word or phrase in the image, they won't put it: they just don't understand letters and words. They understand just pixels and patterns made of pixels. They give you some weird alien looking letters, but nothing really usable for that.

There are some apps that target other very niche but lucrative markets, such as game assets (sprites, backgrounds, effects, etc). I am waiting to use one of those, because they just launched their open beta, and I am wondering how good of a quality those assets will be, and how likely you are to get a wide range of assets in a consistent style. I imagine that would be their biggest struggle: getting a full game asset set in a consistent style.

Of course, the people who are really good at what their do are safe: AI can't get specific and can't take feedback such as "make its hair longer" or "change its background to be more greenish". You can try to iterate on a prompt, but after all, the AI will give you back what they can, not what you want. You either accept it or try again until you are satisfied.

Writing

In the writing space, I think there is mostly ChatGPT, OpenAI's creation, that is the biggest player. Even though it can not solve complex math problems or generate flawless code, it can still be useful in some areas that require creative writing.

copywriting: This could speed up MVP development for startups and marketing website creation. But professionals still are needed because crafting messages that resonate with specific peronas is an art in itself. But in this space, newbies, beware.

fantasy writing: For some quick background stories for some characters, some filler texts here and there, I think AI can do a good job. But the professional writers are safe, because AI writing resembles more the soul-less corporate writing than anything. It has no personal touch and sometimes contradicts itself pretty easily.

Conclusion

All in all, generative AI is becoming more and more relevant and manages to create better and better content. The more advancement we see in this space, the more difficult it will be for new people to make a living in the creative domains. Beginner copywriters or artists will become obsolete, and will have to perfect their craft either pro-bono or as a hobby. They won't be able to pick up a career in those domains, because the more beginner level art will be generated by AI, and you will have to come with some better-than-the-average-AI skills to compete with them. But those skills require time and effort (and maybe even a career) to develop, and you can't develop them because the same results can be obtained from an AI faster and for cheaper. It's kind of like the egg and the chicken problem.