# Is this chart lying to me? Automating the detection of misleading visualizations

## TL;DR

This paper introduces Misviz, a large dataset for detecting misleading visualizations, and evaluates models on their ability to identify design violations that distort data interpretation.

## Contribution

The work provides the first large, annotated datasets for training and evaluating AI models to detect misleading visualizations and their design rule violations.

## Key findings

- State-of-the-art models still struggle with the detection task.
- The datasets enable comprehensive evaluation of AI and rule-based systems.
- Detection of misleading visualizations remains a challenging problem.

## Abstract

Misleading visualizations are a potent driver of misinformation on social media and the web. By violating chart design principles, they distort data and lead readers to draw inaccurate conclusions. Prior work has shown that both humans and multimodal large language models (MLLMs) are frequently deceived by such visualizations. Automatically detecting misleading visualizations and identifying the specific design rules they violate could help protect readers and reduce the spread of misinformation. However, the training and evaluation of AI models has been limited by the absence of large, diverse, and openly available datasets. In this work, we introduce Misviz, a benchmark of 2,604 real-world visualizations annotated with 12 types of misleaders. To support model training, we also create Misviz-synth, a synthetic dataset of 57,665 visualizations generated using Matplotlib and based on real-world data tables. We perform a comprehensive evaluation on both datasets using state-of-the-art MLLMs, rule-based systems, and image-axis classifiers. Our results reveal that the task remains highly challenging. We release Misviz, Misviz-synth, and the accompanying code.