AI or Not

AI or Not (aiornot.com) is an AI image detection tool that analyzes photographs to determine the probability that an image was generated by an artificial intelligence system such as Midjourney, DALL-E, Stable Diffusion, or other generative AI models, rather than captured by a camera.

For OSINT investigators, AI image detection is increasingly critical as generative AI creates synthetic media that is visually indistinguishable from genuine photographs. Synthetic images are actively used in disinformation campaigns, fraud schemes, romance scams, and fake identity creation — investigative contexts where distinguishing AI-generated images from real photographs directly affects investigative conclusions.

When a suspect profile picture, documentary photograph, or claimed evidence image is discovered during an investigation, running it through AI or Not provides a probability assessment of synthetic origin. A high AI probability score warrants deeper verification — cross-searching the image through reverse image search, facial recognition databases, and examining EXIF data to identify inconsistencies.

For romance fraud investigations, where fraudulent profiles frequently use synthetic or stolen profile photos, AI detection serves as an efficient first-pass filter. AI-generated profile pictures often exhibit characteristic artifacts — perfect skin texture, symmetric features, unusual background inconsistencies, or ear and jewelry rendering errors — that AI detection tools are trained to identify.

In disinformation and influence operation investigations, AI image detection helps distinguish between authentic documentary evidence and fabricated imagery. The ability to identify synthetic imagery is essential for maintaining evidentiary integrity and avoiding being misled by manipulated visual evidence.

AI or Not processes images through computer vision models trained on confirmed AI-generated and real photograph datasets. Its outputs are probabilistic rather than definitive — a high AI probability score indicates likelihood, not certainty. False positives (real photos classified as AI) and false negatives (AI images classified as real) occur.

Document the image analyzed (with hash), the AI probability score returned, and the query timestamp. Always corroborate AI detection findings with additional verification methods before drawing firm conclusions.