You may not recognise this ‘painting’ of a handsome bass player in his early 30’s. It is an odd mixture of accuracy (eg. the Roland label on the back of a keyboard) and strange twists (‘Yainara’ rather than ‘Yamaha’ or the way the six string bass becomes a four string on the other side of the mic stand). In fact, if you’ve been keeping an eye on digital art trends in the last year or so, you will probably recognise it as an AI-generated image.

I produced it using krea.ai, a tool I recently came across which offers free image enhancement of photos. It is a fascinating picture, despite the flaws, although it isn’t very recognisable as me. Mind you, when you look at the source picture I fed in, perhaps that isn’t entirely surprising:

Given the motion blur, I don’t think you’d have much hope of recognising me with that one either. Is that scary or exciting? I think it certainly fits somewhere in the region of fascinating. If I was an artist who relied on producing images of people that would mainly be printed or delivered digitally I’d be worried though. I expect that it would be possible to feed a suitable AI tool other pictures of me from the same period and generate something with a much better likeness. It wasn’t instant (as you will find out if you try it yourself) but it was a fraction of the time it would take to set up a canvas and get the lids off the necessary paints, let alone the time needed to do the painting.