A few years ago I shot a nice picture of two of the younger members of the family jumping over small waves at the seaside. It’s a lovely picture showing the innocence and joy of youth and, though I say so myself, the photograph was perfectly timed with both children in the air at the same time. I guess that they would have been about four years old when it was taken and we have a print of it framed in the house. Fast-forward another three or four years and one of the kids in the picture asked me if it had been Photoshopped. I was taken aback and reminded him that this is what I do for a living; taking good photographs without the need to resort to fabrication. It then struck me that they have lived all of their lives in a world where pretty much every image they see is subjected to some sort of manipulation and that their trust in the accuracy of literally anything is pretty limited.

The advent of artificial intelligence in photography has taken the reasons to mistrust what we see to a whole new level and, rather than go through the reasoning again, I would invite you to read a piece written by my colleague Andrew Wiard about the dangers of generative AI now that Adobe Photoshop has joined the long list of applications that allow you to manipulate or create pictures based on a few words of text. You can find his excellent piece here.

The picture that was used to go with his article and the one you see at the top of this post were created by me, using text prompts in Adobe Photoshop to create and image based on an existing photo from Kew Gardens of some giant Lilly pads. There are a few things that I need to emphasise here:

• Before this I had never used generative AI
• It was altered using text prompts alone
• There was no ‘traditional image manipulation’ used
• It took around five minutes from start to finish

The finished image isn’t a work of art and I don’t claim any credit for what the software produced. What I do claim is that this is absolute proof that it takes very little skill to produce something that could be used to fool people. Anyone reading this who hasn’t seen the AI images of Pope Francis wearing a white puffer jacket or sitting astride a large motorcycle should Google them. They are amazingly good and almost believable.

In my world having so many believable fakes doing the rounds causes all sorts of issues. The main one is that real pictures are treated with the same amount of suspicion as the fakes and that makes the world a worse place. Andrew’s article on The BPPA’s website makes all of the points that need to be made and I am right behind any and all initiatives to find a way of labelling fakes and verifying real images. AI is fun but we need to make sure that it doesn’t destroy photography.