Bulk Image Scanner for Gibberish Text, Watermarks, Signatures, etc

I am looking for an expert Python programmer who can create a script that can efficiently scan bulk JPEG and PNG images and flag any that contain gibberish text, watermarks, or signatures (the type of text output that often occurs within AI generated art). The script will need to conduct feature recognition (OCR) and accurately discover hidden or unclear marks. In the past I briefly tried using PyTesseract, but I wasn't able to come up with a working solution. Perhaps this task is more difficult considering the "text" I'm trying to find is often very very small and consists of gibberish text, cursive, or numbers, but I'm hoping it should still be possible. I've attached some sample images so you can see what we'd like the script to be able to flag. We don't need to pull any text or anything like that from the images, we just need to be able to flag/filter any images that include text/symbols/etc when bulk processing. --- Key aspects of this project include: Output Format: - Maintain the original format for the images that are flagged. I do not need any conversion done, just the flag for my attention. Compilation: - I prefer flagged images to be tagged using the Tags function in macOS. However, I am open to other effective solutions you might suggest. Ideal Skills: - Proficient in Python scripting - Experience in image processing - Knowledge of macOS system functions.