by JohnM » Thu Sep 05, 2024 3:16 pm
As a follow-on to my previous dialogue with ChatGPT about image assessment, I asked it how it calculated a numerical score for an image. As the response included factors such as "emotion" and "storytelling", I then asked how an AI can assess these criteria. I then asked the bot to analyse the cover image for Issue 01 2024-25 of the newsletter, before realising that its assessment criteria had changed since the beginning of the dialogue, so I asked it why, and whether the image would score differently against the "old" criteria.
You do not have the required permissions to view the files attached to this post.