Artists Take Heed – AI is Coming!

This is DALL-E. It’s a deep learning model developed by OpenAI to generate digital images from natural language descriptions, or prompts. These images on screen were entirely created by AI, and in this video, I will test it out by writing some of my own prompts to show you what it can do.

“A painting of an unknown medieval man with a hoe”
Each prompt produces four square image costing a total of one credit. (They give you 50 free credits when you sign up, and 15 free credits every month – you can buy more credits at about 2 cents per image for the highest quality.

“The official photographic portrait of the future president of the United States”
Okay, so the photo quality isn’t that great, but it looks like the AI is predicting that Donald Trump is going to reclaim the presidency, or at least, somebody who looks like him.

“A van Gogh style painting of an Australian rugby player”
Yeah, it looks all right, I suppose.

“3D render of a fruit bowl with fruit made of balloons in a violet room”

And now a quote from Al Gore from the recent World Economic Forum, “Boiling oceans caused by the quivalent of 600,000 Hiroshima-class atomic bombs exploding every single day on the Earth”

“Evil dictator forcing medical treatment on people”
“It looks like this request may not follow our content policy”. So, not everything is drawable based on their content policy. “Do not attempt to create, upload, or share images that are not G-rated or that could cause harm”. Fair enough. Also, they state, “Don’t mislead your audience about AI involvement”, and “Do not upload images of people without their consent”. That’s interesting. Why would I be uploading pictures?

Well, it turns out the AI can modify existing photos. For example, I uploaded this photo of an elderly gentleman walking (don’t worry, I got his permission), and then you can crop a square part of the image, and then you can Edit or Generate a variation. Let’s try generating a variation first. Okay, that’s kind of cool, but obviously, that would never pass as a real photo. Now we’ll edit the image. You just have to erase part of the image and write a description of what you want to be there. In this case, “A man walking a large robot dog”. And here’s the result.

Here’s another photo of a man pumping his fist. Here he is carrying a flaming torch. Here he is carrying a Chinese flag. Here he is carrying a pink Teddy bear. And here he is eating the top off an ice cream cone.

You can also hit the “Surprise Me” button: “Panda mad scientist mixing sparkling chemicals, digital art”.
“A bowl of soup that looks like a monster, knitted out of wool”. And, “an oil painting by Matisse of a humanoid robot playing chess”.

So what do you think? Will AI take over? Is art by humans dead?

MUSIC
Allégro by Emmit Fenn