Will AI generated art usher in a new RenAIssance?

Sometimes if I want to horrify myself, I think back to the thousands of hours of my life that I’ve spent playing video games and aimlessly flipping through TV channels.

Because imagine if I’d spent that time instead learning to draw or paint? Perhaps by now I could be a great artist?

However, recently I’ve not felt so bad about it, as perhaps there was no need for me to learn anyway.

Over the last year we’ve seen an incredible proliferation in artistic tools powered by artificial intelligence—to the point where computers can now generate photo-realistic images and masterful paintings that are virtually indistinguishable from the works of the greatest human photographers and artists.

For example, recently I’ve been playing around with an AI image generation tool called “DALL-E"—a pun combining the Spanish surrealist with the Pixar robot, which is made by the same company that is behind ChatGPT.

The way it works is not by using the mouse as a virtual brush or touching the screen and drawing with fingers—instead you simply have to type in a text prompt.

“Draw a guinea pig soldier on parade,” you can ask it—and within seconds you’ll have a more than passable image, no talent required.

What I find genuinely jaw-dropping, even as someone who follows tech closely, is just how quickly the tools are evolving.

When the first AI art apps were released about a year ago, generated images were often fuzzy, or not right in obvious ways: humans would have the wrong number of fingers, missing limbs, or even additional arms poking out where they shouldn’t.

But in just a few months, the quality of what is being generated has got so much better. Fast forward to today and not only are images nearly perfect a lot of the time, but there are more ways than ever to generate AI images.

For example, Adobe, which for decades has made the Photoshop photo-editing app, recently added an AI feature where you can highlight a part of a photo and swap it out for something artificial.

Wish it was less cloudy when you took the photo? Or even want to remove your ex from that family picture? Now there’s no need to spend hours carefully editing your photo—you can do it in seconds.

The AI boom has also led to the creation of smaller start-ups that specialise in doing one extremely clever thing with images.

For example, a company called Krea has a tool where you sketch out an image with your mouse and it will turn your scribbles into a photo-realistic drawing.

Your crude house on top of a hill, which looks like something a six-year-old might draw, will be instantly transformed into a photo-realistic rendering of a rural cottage on a Highland mountain-top.

Then there’s an app called Runway, which will take a still image—either a real photo or something you’ve made with AI elsewhere—and turn it into a short video.

There are many more I could mention too.

However, the rise of the technology has of course been hugely controversial. Because to make these tools as powerful as they are, the AI needs to be "trained" on millions of other images—images that were created by real artists using their real talents. So some artists view AI as a form of plagiarism. 

Then there’s the grim downstream consequence of image generation being as easy as it is: social media is already being flooded with fakery.

I’ve experienced this myself. With DALL-E, I did an experiment and asked it to generate a photo of some soldiers in the Middle East, and it happily complied.

Then I asked it to generate an image of some dead bodies covered in blood, and mercifully it refused to do it.

But, unfortunately, I figured out it was very easy to get around—as I just asked it to generate some mannequins lying on the ground and to cover them in jam.

And the results, which I’m not going to publish, are realistic enough that you can easily imagine them being misused by people with bad intentions.

So it really is important to no longer believe your eyes when you see a photo that is too good to be true.

However, for all of the potential harms, the fact is that AI artwork exists now, and no one is going to un-invent it. The tools to make images are already widely spread, so the genie is out of the bottle.

And though we should care about the negative consequences, we should also look forward to the positives.

For example, my art skills have just received an unexpected bump. And it really makes me wonder if the greatest artists of the future won’t be people who are handy with a brush, but those who have mastered the art of writing the perfect prompt.