# How good are humans at detecting AI-generated images? Learnings from an experiment

arXiv: https://arxiv.org/abs/2507.18640 · HTML: https://mlsanity.com/paper/2507.18640/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- Builds on 3 of its 7 references that are in the map.
- First posted on arXiv 12 May 2025.

## Builds on (most cited first, 3 of 3)

- [Generative Adversarial Networks](https://mlsanity.com/paper/2203.00667/) (2022)
- [Using Human Feedback to Fine-tune Diffusion Models without Any Reward Model](https://mlsanity.com/paper/2311.13231/) (2023)
- [Seeing is not always believing: Benchmarking Human and Model Perception of AI-Generated Images](https://mlsanity.com/paper/2304.13023/) (2023)

## Built on by (most cited first, 7 of 7)

- [Gram-MMD: A Texture-Aware Metric for Image Realism Assessment](https://mlsanity.com/paper/2604.03064/) (2026)
- [The Unwinnable Arms Race of AI Image Detection](https://mlsanity.com/paper/2509.21135/) (2025)
- [We are not able to identify AI-generated images](https://mlsanity.com/paper/2512.22236/) (2025)
- [From AI-Generated Content to Agentic Action: Security and Safety Threats in Generative AI](https://mlsanity.com/paper/2605.16471/) (2026)
- [The CIFAR Synthetic Evidence Corpus for Detecting AI-Generated Evidence](https://mlsanity.com/paper/2606.07916/) (2026)
- [Generative AI Literacy Training Improves Intelligence Analysts' Discrimination of Real and AI-Generated Images](https://mlsanity.com/paper/2606.28510/) (2026)
- [A Benchmark & Dataset for Detecting AI-Manipulated Visual Evidence in the Court System](https://mlsanity.com/paper/2609.37783/) (2026)

## Abstract (arXiv)

As AI-powered image generation improves, a key question is how well human beings can differentiate between "real" and AI-generated or modified images. Using data collected from the online game "Real or Not Quiz.", this study investigates how effectively people can distinguish AI-generated images from real ones. Participants viewed a randomized set of real and AI-generated images, aiming to identify their authenticity. Analysis of approximately 287,000 image evaluations by over 12,500 global participants revealed an overall success rate of only 62%, indicating a modest ability, slightly above chance. Participants were most accurate with human portraits but struggled significantly with natural and urban landscapes. These results highlight the inherent challenge humans face in distinguishing AI-generated visual content, particularly images without obvious artifacts or stylistic cues. This study stresses the need for transparency tools, such as watermarks and robust AI detection tools to mitigate the risks of misinformation arising from AI-generated content