Drop a photo here, or click to upload and we will check for digital enhancements and hidden AI patterns. Our advanced scanner looks for the subtle tells that the human eye often misses.
Click to upload or drag and drop
JPG, PNG, or WEBP (max. 3MB) — or press Ctrl+V to paste
Our scanner examines pixel-level patterns, edge inconsistencies, and unnatural texture repetition that AI image generators leave behind. These artifacts are often imperceptible to the human eye but statistically significant.
We apply a Fast Fourier Transform (FFT) to decompose the image into frequency components. AI-generated images tend to have distinct spectral signatures in the high-frequency bands that differ from photographs taken with a real camera.
A large multimodal vision model evaluates the image holistically, looking for semantic inconsistencies such as impossible lighting, anatomically incorrect hands, font errors, and warped backgrounds — common failure modes for generative AI.
Generative AI models such as Midjourney, Stable Diffusion, DALL·E, and Adobe Firefly can produce photorealistic imagery in seconds. While these tools have enormous creative value, they are also increasingly used to create misinformation, fabricate evidence, and deceive audiences at scale. Being able to distinguish a real photograph from a synthetic image has become a critical digital literacy skill.
AI Detective combines multiple analysis techniques — pixel-level artifact detection, frequency domain inspection, and large language model vision analysis — to provide a comprehensive assessment of whether an image is likely to have been generated by an AI model. No single technique is foolproof, which is why we layer several independent signals and present them transparently in the results.
The tool is free to use and designed for journalists, educators, researchers, and anyone who wants a second opinion before sharing or publishing an image. Always exercise critical judgment alongside automated analysis.