Semi-Truths: A Large-Scale Dataset of AI-Augmented Images for Evaluating Robustness of AI-Generated Image detectors

TL;DR

SEMI-TRUTHS is a large-scale dataset of real and AI-augmented images designed to evaluate the robustness of AI-generated image detectors across various perturbations and data distributions.

Contribution

We introduce SEMI-TRUTHS, a comprehensive dataset with diverse augmentations and metadata for standardized evaluation of detector robustness.

Findings

Detectors show varying sensitivities to different perturbations.

Performance depends on augmentation types and data distributions.

Insights reveal limitations and biases in current detection methods.

Abstract

Text-to-image diffusion models have impactful applications in art, design, and entertainment, yet these technologies also pose significant risks by enabling the creation and dissemination of misinformation. Although recent advancements have produced AI-generated image detectors that claim robustness against various augmentations, their true effectiveness remains uncertain. Do these detectors reliably identify images with different levels of augmentation? Are they biased toward specific scenes or data distributions? To investigate, we introduce SEMI-TRUTHS, featuring 27,600 real images, 223,400 masks, and 1,472,700 AI-augmented images that feature targeted and localized perturbations produced using diverse augmentation techniques, diffusion models, and data distributions. Each augmented image is accompanied by metadata for standardized and targeted evaluation of detector robustness. Our findings suggest that state-of-the-art detectors exhibit varying sensitivities to the types and degrees of perturbations, data distributions, and augmentation methods used, offering new insights into their performance and limitations. The code for the augmentation and evaluation pipeline is available at https://github.com/J-Kruk/SemiTruths.