Artificial Artists

By Lucy Hattersley. Posted

Artificial Intelligence is still in the news, and for good reason. There are incredible strides appearing in the world of computer-generated images, video and text. Generative art and video is based upon an open-source technology called Stable Diffusion, which can create any artwork you desire based on text prompts. You can test it out online at stablediffusionweb.com if you want to paint the cheese moon landing, or the distracted boyfriend meme by Picasso. The open-source Stable Diffusion 2 GitHub code is found here.

If you think artists have it bad, then spare a thought for your lonely writers. We’re staring down the barrel of ChatGPT and Google’s response, Bard.

The creative mill

As Rob pointed out last month, ChatGPT’s copy is generously described as ‘hit-and-miss’. A more accurate assessment would be: “dumber than a tall person hopping at the very peak of Dunning-Kruger’s mount stupid”. It is convincing though and a lot of online copy is equally misguided. Which may be the point.

Getty Images is suing Stable Diffusion for $1.8-trillion because: “It is Getty Images’ position that Stability AI unlawfully copied and processed millions of images protected by copyright.”

Beyond copyright and clarity, there are legitimate ethical concerns. Since last month tech wizards with flexible morals have already got around safeguards by getting ChatCPT to role-play as a different AI called ‘DAN’ which stands for ‘Do-anything-Now’.

AI can create anything as dark as the mind can conjure. A quick traipse through the internet will reveal some very dark thoughts indeed.

Positive thoughts

It’s easy to knee-jerk against this advancement. I love what I do, but technology should replace people. I can quite happily sit on a beach and read poetry to my friends while computers do all the work.

The IBM/Jim Henson’s Machines Should Work; People should think speech from 1967 comes to mind. Computers are supposed to take out the trash, clean the floor, do the dishes and the washing up and we (the humans) are supposed to sit on the beach and share our stories, make up music, and paint pictures.

The future of machine-learning technology is, I suspect, in more mundane, but ultimately useful, applications. The robot vacuum cleaner that picks up the socks from your ‘floordrobe’ and puts them back in a drawer (or in the wash) is ultimately far more valuable to us than one that paints selfies.

Of course, that machine would probably need wheels and arms and a mechanical nose alongside the AI smarts to know what scruffy socks look like. Oh and look, scientists from Tel Aviv are working on a robot with a biological nose. Of course, it has Raspberry Pi inside.

This is the real stuff of the future. Where hardware and software and robotics and AI-smarts come together to do useful stuff while we all chill.

From The MagPi store

Subscribe

Subscribe to the newsletter

Get every issue delivered directly to your inbox and keep up to date with the latest news, offers, events, and more.