# EviScene: Beyond "Is It AI?" for Agentic Image Factuality Assessment

arXiv: https://arxiv.org/abs/2610.14609 · HTML: https://mlsanity.com/paper/2610.14609/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Multimodal and embodied agents, one of 16 directions in LLM agents and tool use.
- Builds on 29 of its 56 references that are in the map.
- First posted on arXiv 10 Oct 2026.

## Placed in

- LLM agents and tool use / [Multimodal and embodied agents](https://mlsanity.com/agents/multimodal-and-embodied-agents/)

## Builds on (most cited first, 29 of 29)

- [ReAct: Synergizing Reasoning and Acting in Language Models](https://mlsanity.com/paper/2210.03629/) (2022)
- [On-Policy Distillation of Language Models: Learning from Self-Generated Mistakes](https://mlsanity.com/paper/2306.13649/) (2023)
- [SWIFT:A Scalable lightWeight Infrastructure for Fine-Tuning](https://mlsanity.com/paper/2408.05517/) (2024)
- [AgentTuning: Enabling Generalized Agent Abilities for LLMs](https://mlsanity.com/paper/2310.12823/) (2023)
- [GenImage: A Million-Scale Benchmark for Detecting AI-Generated Image](https://mlsanity.com/paper/2306.08571/) (2023)
- [WISE: A World Knowledge-Informed Semantic Evaluation for Text-to-Image Generation](https://mlsanity.com/paper/2503.07265/) (2025)
- [NewsCLIPpings: Automatic Generation of Out-of-Context Multimodal Media](https://mlsanity.com/paper/2104.05893/) (2021)
- [End-to-End Multimodal Fact-Checking and Explanation Generation: A Challenging Dataset and Models](https://mlsanity.com/paper/2205.12487/) (2022)
- [Breaking Common Sense: WHOOPS! A Vision-and-Language Benchmark of Synthetic and Compositional Images](https://mlsanity.com/paper/2303.07274/) (2023)
- [SNIFFER: Multimodal Large Language Model for Explainable Out-of-Context Misinformation Detection](https://mlsanity.com/paper/2403.03170/) (2024)
- [Dual Data Alignment Makes AI-Generated Image Detector Easier Generalizable](https://mlsanity.com/paper/2505.14359/) (2025)
- [Multimodal Automated Fact-Checking: A Survey](https://mlsanity.com/paper/2305.13507/) (2023)
- [DEFAME: Dynamic Evidence-based FAct-checking with Multimodal Experts](https://mlsanity.com/paper/2412.10510/) (2024)
- [VERITE: A Robust Benchmark for Multimodal Misinformation Detection Accounting for Unimodal Bias](https://mlsanity.com/paper/2304.14133/) (2023)
- [AVerImaTeC: A Dataset for Automatic Verification of Image-Text Claims with Evidence from the Web](https://mlsanity.com/paper/2505.17978/) (2025)
- [FACTIFY3M: A Benchmark for Multimodal Fact Verification with Explainability through 5W Question-Answering](https://mlsanity.com/paper/2306.05523/) (2023)
- [Evaluating Image Hallucination in Text-to-Image Generation with Question-Answering](https://mlsanity.com/paper/2409.12784/) (2024)
- [T2I-FactualBench: Benchmarking the Factuality of Text-to-Image Models with Knowledge-Intensive Concepts](https://mlsanity.com/paper/2412.04300/) (2024)
- [VisBrowse-Bench: Benchmarking Visual-Native Search for Multimodal Browsing Agents](https://mlsanity.com/paper/2603.16289/) (2026)
- [Beyond Artifacts: Real-Centric Envelope Modeling for Reliable AI-Generated Image Detection](https://mlsanity.com/paper/2512.20937/) (2025)
- [MIRROR: Manifold Ideal Reference ReconstructOR for Generalizable AI-Generated Image Detection](https://mlsanity.com/paper/2602.02222/) (2026)
- [Semantic Visual Anomaly Detection and Reasoning in AI-Generated Images](https://mlsanity.com/paper/2510.10231/) (2025)
- [Towards On-Policy Data Evolution for Visual-Native Multimodal Deep Search Agents](https://mlsanity.com/paper/2605.10832/) (2026)
- [Visual-Seeker: Towards Visual-Native Multimodal Agentic Search via Active Visual Reasoning](https://mlsanity.com/paper/2606.15231/) (2026)
- [VeriTaS: The First Dynamic Benchmark for Multimodal Automated Fact-Checking](https://mlsanity.com/paper/2601.08611/) (2026)
- ["Image, Tell me your story!" Predicting the original meta-context of visual misinformation](https://mlsanity.com/paper/2408.09939/) (2024)
- [RW-Post: Auditable Evidence-Grounded Multimodal Fact-Checking in the Wild](https://mlsanity.com/paper/2605.10357/) (2026)
- [Rethinking the Readout: Unlocking Video Backbones for AI-Generated Video Detection](https://mlsanity.com/paper/2607.15321/) (2026)
- [FAGER: Factually Grounded Evaluation and Refinement of Text-to-Image Models](https://mlsanity.com/paper/2605.19111/) (2026)

## Abstract (arXiv)

Determining whether an image is AI-generated does not establish whether the situation it depicts is factually supported. We formulate Image Factuality Assessment: given an image and a generic verification instruction, a system returns a supported or refuted verdict with an evidence-grounded explanation. Without a supplied textual claim, the system must progressively refine factual questions during investigation and align external evidence with the specific situation depicted. We introduce EviLens, an expert-reviewed benchmark of 10,174 supported and refuted images from controlled construction and web collection, covering twelve factual error types and evaluating both verdict accuracy and evidence sufficiency. We develop EviScene, a multimodal ReAct agent that interleaves visual inspection, external retrieval, and evidence analysis, using new findings to refine its questions and revisit relevant image details. We train a compact policy with supervised fine-tuning on reviewed teacher trajectories, followed by privileged on-policy self-distillation applied to verified decisions from the policy's own investigations. On the 1,527-image test set, the trained Qwen3.5-9B agent achieves 79.05% balanced accuracy and a Strict Evidence Sufficiency Rate (SESR) of 52.26%, improving over the same-backbone base agent by 6.71 and 9.37 percentage points, respectively. Its SESR is the highest among the evaluated agents.Code, models, and data will be released at https://github.com/wangzian828/EVISCENE and https://huggingface.co/ZiAnwang/EviScene.