Related papers: Learning Human-Perceived Fakeness in AI-Generated Videos via Multimodal LLMs

Learning Human-Perceived Fakeness in AI-Generated Videos via Multimodal LLMs

URL: http://arxiv.org/abs/2509.22646v2
Date: Wed, 01 Oct 2025 05:14:25 GMT
Title: Learning Human-Perceived Fakeness in AI-Generated Videos via Multimodal LLMs
Authors: Xingyu Fu, Siyi Liu, Yinuo Xu, Pan Lu, Guangqiuse Hu, Tianbo Yang, Taran Anantasagar, Christopher Shen, Yikai Mao, Yuanzhe Liu, Keyush Shah, Chung Un Lee, Yejin Choi, James Zou, Dan Roth, Chris Callison-Burch,
Abstract summary: We introduce DeeptraceReward, a benchmark that annotates human-perceived fake traces for video generation reward.<n>The dataset comprises 4.3K detailed annotations across 3.3 high-quality generated videos.<n>We consolidate these annotations into 9 major categories of deepfake traces that lead humans to identify a video as AI-generated.
Score: 92.02382309654263
License: http://creativecommons.org/licenses/by-nc-sa/4.0/
Abstract: Can humans identify AI-generated (fake) videos and provide grounded reasons? While video generation models have advanced rapidly, a critical dimension -- whether humans can detect deepfake traces within a generated video, i.e., spatiotemporal grounded visual artifacts that reveal a video as machine generated -- has been largely overlooked. We introduce DeeptraceReward, the first fine-grained, spatially- and temporally- aware benchmark that annotates human-perceived fake traces for video generation reward. The dataset comprises 4.3K detailed annotations across 3.3K high-quality generated videos. Each annotation provides a natural-language explanation, pinpoints a bounding-box region containing the perceived trace, and marks precise onset and offset timestamps. We consolidate these annotations into 9 major categories of deepfake traces that lead humans to identify a video as AI-generated, and train multimodal language models (LMs) as reward models to mimic human judgments and localizations. On DeeptraceReward, our 7B reward model outperforms GPT-5 by 34.7% on average across fake clue identification, grounding, and explanation. Interestingly, we observe a consistent difficulty gradient: binary fake v.s. real classification is substantially easier than fine-grained deepfake trace detection; within the latter, performance degrades from natural language explanations (easiest), to spatial grounding, to temporal labeling (hardest). By foregrounding human-perceived deepfake traces, DeeptraceReward provides a rigorous testbed and training signal for socially aware and trustworthy video generation.

Related papers

Skyra: AI-Generated Video Detection via Grounded Artifact Reasoning [66.51617619673587]
We present Skyra, a specialized large language model (MLLM) that identifies human-perceivable visual artifacts in AI-generated videos.<n>To support this objective, we construct ViF-CoT-4K for Supervised Fine-Tuning (SFT), which represents the first large-scale AI-generated video dataset with fine-grained human annotations.<n>We then develop a two-stage training strategy that systematically enhances our model's artifact's-temporal perception, explanation capability, and detection accuracy.
arXiv Detail & Related papers (2025-12-17T18:48:26Z)
Leveraging Pre-Trained Visual Models for AI-Generated Video Detection [54.88903878778194]
The field of video generation has advanced beyond DeepFakes, creating an urgent need for methods capable of detecting AI-generated videos with generic content.<n>We propose a novel approach that leverages pre-trained visual models to distinguish between real and generated videos.<n>Our method achieves high detection accuracy, above 90% on average, underscoring its effectiveness.
arXiv Detail & Related papers (2025-07-17T15:36:39Z)
BrokenVideos: A Benchmark Dataset for Fine-Grained Artifact Localization in AI-Generated Videos [63.03271511550633]
BrokenVideos is a benchmark dataset of 3,254 AI-generated videos with meticulously annotated, pixel-level masks highlighting regions of visual corruption.<n>Our experiments show that training state of the art artifact detection models and multi modal large language models (MLLMs) on BrokenVideos significantly improves their ability to localize corrupted regions.
arXiv Detail & Related papers (2025-06-25T03:30:04Z)
ExDDV: A New Dataset for Explainable Deepfake Detection in Video [23.169975307069066]
We introduce ExDDV, the first dataset and benchmark for Explainable Deepfake Detection in Video.<n>We evaluate a number of vision-language models on ExDDV, performing experiments with various fine-tuning and in-context learning strategies.<n>Our results show that text and click supervision are both required to develop robust explainable models for deepfake videos.
arXiv Detail & Related papers (2025-03-18T16:55:07Z)
Vulnerability-Aware Spatio-Temporal Learning for Generalizable Deepfake Video Detection [14.586314545834934]
We propose a fine-grained deepfake video detection approach called FakeSTormer.<n>Specifically, we introduce a multi-task learning framework that incorporates two auxiliary branches for explicitly attending artifact-prone spatial and temporal regions.<n>We also propose a video-level synthesis strategy that generates pseudo-fake videos with subtle-temporal artifacts.
arXiv Detail & Related papers (2025-01-02T10:21:34Z)
What Matters in Detecting AI-Generated Videos like Sora? [51.05034165599385]
Gap between synthetic and real-world videos remains under-explored. In this study, we compare real-world videos with those generated by a state-of-the-art AI model, Stable Video Diffusion. Our model is capable of detecting videos generated by Sora with high accuracy, even without exposure to any Sora videos during training.
arXiv Detail & Related papers (2024-06-27T23:03:58Z)
Unmasking Illusions: Understanding Human Perception of Audiovisual Deepfakes [49.81915942821647]
This paper aims to evaluate the human ability to discern deepfake videos through a subjective study. We present our findings by comparing human observers to five state-ofthe-art audiovisual deepfake detection models. We found that all AI models performed better than humans when evaluated on the same 40 videos.
arXiv Detail & Related papers (2024-05-07T07:57:15Z)
Undercover Deepfakes: Detecting Fake Segments in Videos [1.2609216345578933]
deepfake generation is a new paradigm of deepfakes which are mostly real videos altered slightly to distort the truth. In this paper, we present a deepfake detection method that can address this issue by performing deepfake prediction at the frame and video levels. In particular, the paradigm we address will form a powerful tool for the moderation of deepfakes, where human oversight can be better targeted to the parts of videos suspected of being deepfakes.
arXiv Detail & Related papers (2023-05-11T04:43:10Z)
Copy Motion From One to Another: Fake Motion Video Generation [53.676020148034034]
A compelling application of artificial intelligence is to generate a video of a target person performing arbitrary desired motion. Current methods typically employ GANs with a L2 loss to assess the authenticity of the generated videos. We propose a theoretically motivated Gromov-Wasserstein loss that facilitates learning the mapping from a pose to a foreground image. Our method is able to generate realistic target person videos, faithfully copying complex motions from a source person.
arXiv Detail & Related papers (2022-05-03T08:45:22Z)
What's wrong with this video? Comparing Explainers for Deepfake Detection [13.089182408360221]
Deepfakes are computer manipulated videos where the face of an individual has been replaced with that of another. In this work we develop, extend and compare white-box, black-box and model-specific techniques for explaining the labelling of real and fake videos. In particular, we adapt SHAP, GradCAM and self-attention models to the task of explaining the predictions of state-of-the-art detectors based on EfficientNet.
arXiv Detail & Related papers (2021-05-12T18:44:39Z)

This list is automatically generated from the titles and abstracts of the papers in this site.