AI or Not, developed by tech entrepreneurs based in Silicon Valley, is an innovative platform that detects whether media files like images, videos, and audio are AI-generated. Originally focused on identifying fake NFTs, the tool has expanded its scope to address broader societal issues such as misinformation, deepfakes, and their potential to influence elections and fuel propaganda, particularly during crises like the ongoing wars. The platform’s ability to detect AI-manipulated content, including military-related fake images, plays a crucial role in combating disinformation in times of conflict and political turbulence. AI or Not offers free and paid subscription plans for individuals, and provides fee-based services to entities.

Theory of Change

Providing accessible tools for the detection of fabricated content helps protect democratic processes, foster transparency, and ensure informed decision-making in an era of AI-driven disinformation.

Activities

AI or Not uses advanced algorithms to scan and analyze media files, detecting inconsistencies such as digital footprints, pixel patterns, and metadata alterations that indicate the involvement of AI. The system compares these against datasets of known AI-generated content, enabling it to identify manipulated elements. The final product delivered to users includes a detailed report that breaks down the analysis. This report highlights whether the image was AI-generated, sections of the image where AI alterations were detected, information about the origins and digital history of the media, and a confidence score that indicates how certain the tool is about the detection of AI involvement.

Results

The tool has analyzed millions of media assets and detected over 100 million fake NFTs, it is trusted by 200k+ clients worldwide and continues to evolve, expanding its detection capabilities to include deepfake videos and AI-generated voice recordings. As one example, the platform analyzed AI-generated images that depicted Taylor Swift endorsing Donald Trump. While it was fairly obvious they were fake, AI or Not was able to show the precise AI used to produce the fake.