The Slop Sense web service has been launched to test AI image detection skills

Developers presented Slop Sense — an interactive web tool designed to test users' ability to distinguish real photos from images created by generative models. The service's mechanics are based on a time game: the user is given exactly 60 seconds to classify a series of visual content. Points are awarded for each correct answer, and additional points are provided for a series of correct decisions. The project aims to demonstrate the current level of realism of generative content and the complexity of its verification by the human eye. The tool is available in the browser without the need to install additional software. Statistics and discussion of the project are available in the Hacker News thread, where users share the results of their attempts and analyze typical artifacts that reveal the work of neural networks. Similar projects are becoming part of the infrastructure for evaluating the quality of visual data in the context of the growing volume of synthetic content. The tool allows you to practically assess how effectively modern models imitate textures, lighting, and details of real objects, which is an important aspect in cybersecurity and media literacy tasks.