AI-Generated Images: What Humans and Machines See When They Look at the Same Image

2026-05-07

AI Summary

This study addresses the growing threat of generative AI–enabled disinformation by developing a detection system that provides human-interpretable explanations for identifying synthetic images. The authors construct a multi-architecture deep detection model trained on AIText2Image, a large-scale text-to-image generation dataset, and integrate 16 distinct explainable AI (XAI) methods. They propose a novel human-centered XAI evaluation framework, leveraging textual and visual feedback from 100 participants to systematically assess the role of visual-language cues in fake image detection and the alignment between XAI outputs and human cognition. Experimental results demonstrate high detection accuracy while revealing significant disparities among XAI methods in terms of human interpretability, offering critical empirical insights for the design of effective and trustworthy explainable AI systems.

Abstract

The misuse of generative AI in online disinformation campaigns highlights the urgent need for transparent and explainable detection systems. In this work, we investigate how detectors for AI-generated images can be more effective in providing human-understandable explanations for their predictions. To this end, we develop a suite of detectors with various architectures and fine-tuning strategies, trained on our large-scale photorealistic fake image dataset, AIText2Image, and assess their performance on state-of-the-art text-to-image AI generators. We integrate 16 different explainable AI (XAI) methods into our detection framework, and the visual explanations are comprehensively refined and evaluated through a novel approach that prioritizes human understanding of AI-generated images, using both textual and visual responses collected from a survey of 100 participants. This framework offers insights into visual-language cues in fake image detection and into the clarity of XAI methods from a human perspective, measuring the alignment of XAI outputs with human preferences.

Problem

• AI-generated images
• explainable AI
• fake image detection
• human-understandable explanations
• disinformation

Innovation

• Explainable AI (XAI)
• AI-generated image detection
• human-centered evaluation
• visual-language cues
• AIText2Image dataset