top of page
< Back

AI in Image Competitions ~ can we tell?

Artificial intelligence in photography is here, it is part of our ever changing environment, and can do amazing things. I've heard it will be able to mask so accurately that feather edges can show a transparent glow... So - in dealing with image competitions - how can we differentiate images created with a camera and adjusted with a computer from images created solely with AI? How do we maintain fairness in competitions? Well - AI can also be used to determine the amount of AI used in...

Artificial intelligence in photography is here, it is part of our ever changing environment, and can do amazing things. I've heard it will be able to mask so accurately that feather edges can show a transparent glow... So - in dealing with image competitions - how can we differentiate images created with a camera and adjusted with a computer from images created solely with AI? How do we maintain fairness in competitions? Well - AI can also be used to determine the amount of AI used in creating an image! There are now websites that can calculate the amount of AI used in an image - from a low percentage of AI Noise Reduction to "99% likely AI-generated". Here's what Google Gemini AI says about one of these sites - SightEngine: https://sightengine.com/ "Sightengine is widely regarded as one of the most reliable commercial and enterprise-grade tools for detecting AI-generated images. Independent benchmarks (including studies evaluating media detection APIs) frequently rank it near the top for its pixel-level analysis capabilities. Key Reliability Metrics & Strengths High Overall Accuracy: In broad benchmark testing across tens of thousands of real and synthetic images, Sightengine typically scores between 88% and 93%+ accuracy, excelling particularly with major text-to-image generators like Midjourney, DALL-E, Flux, and Stable Diffusion. Low False Positive Rate: One of its biggest advantages for professional use (like content moderation) is its low false-positive rate—usually hovering around 2.5%. This means it rarely misidentifies genuine, high-quality camera photography as AI-generated. Pixel-Level Rather Than Metadata Analysis: Sightengine analyzes the actual pixel structure rather than relying purely on digital watermarks or stripped EXIF metadata. This makes it resilient against users stripping metadata or resizing images." Here is a Sightengine report on an image recently submitted to a regional image competition where AI created images were not allowed: I can foresee that competitions will start running entries through one of these sites to determine eligibility and compliance with AI rules.

ai-in-image-competitions-~-can-we-tell

PPA logo.png

Coastal California Photographers is a non-profit, professional & enthusiast Photography organization committed to advancing photographic excellence through education, inspiration, mentoring, and collaborative learning while maintaining the highest ethical practices and conduct standards. Coastal California Photographers is a proud affiliate of Professional Photographers of California (PPC), Professional Photographers of America (PPA), and Photographic Society of America (PSA). 

ppc-logo.png
PSA.png
bottom of page