With the proliferation of AI slop across social media platforms – 25% of posts on TikTok, according to our findings – we have decided to share our expertise in detecting AI-generated content with other experts, fact checkers, journalists, researchers, and even regular users who may want to learn how to better decipher the content they see across social media.

In this guide, we account for a spectrum of synthetic content that can be referred to as “generative AI imagery,” including both still images and moving images or clips.

This guide is structured as a series of steps a user can follow to assess the likelihood that the media content they see has been generated using AI tools. The steps are divided into four focus areas:

1. Before the Image: AI Telltales;
2. Synthetic Artifacts in AI Imagery;
3. Moving Images/Clips;
4. Digital Provenance.

We recommend consulting the steps in the specified order. For relative assurance that an image or video is likely AI-generated, the content in question should meet more than one of the characteristics that follow.