This photo is of headstones in the old colonial Jewish section of the West Terrace cemetery in Adelaide, Australia.
I am not sure how I feel about this one. I asked Chat GPT to colorize this (me having previously turned it to monochrome in post
🙂), increase the image resolution (generally which it did pretty well) and more specifically to increase the legibility of the script on the closest headstone - particularly the English script which was degraded by erosion of the stone. I made my own interpretation and prompted the AI to adopt the same words. Overall it did a pretty good job. But my main complaints are that the English script is far too legible and "new" looking (also quite artificial) and that the Hebrew script in the second headstone looks like random AI gibberish more than actual Hebrew, at least to me (I do not read it, but I am going by appearance.)
I do not know quite what is worse - making an image so good that it looks so photo real it can fool people or making an image pretty good but "bodging" certain elements of it so that it looks rather fake.
😕 As regards the reality I think I could improve it by better prompts about how "aged" the script should look but in any event its an interesting experiment which also raises interesting questions - for example how much modification by AI is OK. I habitually modify my out of camera shots anyway because I like more painterly, impressionistic, artistic images. Is it Ok to do the same with AI and where is the cut off between merely "correcting" using AI and creating an image containing elements that are so deficient in some respect that its fakeness is obvious. And is it OK to use AI skills to make the same kind of artistic interpretations I make anyway by hand (well, with Lightroom and various other software apps). I have no answers other than perhaps the need to be honest.
PS I also added some of my own edits to the final image in post for a more dramatic result.
