A Vision-Language Dual-View Survey for factual fidelity verification
ACL 2026
ICML 2026 EMM-QA Workshop
1 MBZUAI
2 Renmin University of China
3 Harvard University
† Corresponding author: Yuhan Liu · yuhan.liu@mbzuai.ac.ae
As video generators move from localized edits to fully synthetic, cinematic scenes, artifact-centric detection is no longer enough. This survey reframes AIGC-V detection as factual fidelity verification: whether the events, entities, and physical processes depicted in a video remain consistent with the real world.
Modern AI-generated video cannot be understood through frame artifacts alone. Detection must combine low-level intrinsic cues, spatiotemporal consistency, cross-modal consistency, and language-guided world-level reasoning.
The task is not only to label a clip as fake, but to identify which claim, event, entity, or temporal segment breaks consistency with reality.
The visual side covers intrinsic cues and spatiotemporal consistency inside the clip, while the language side covers cross-modal consistency and language-guided world-level reasoning.
Trustworthy systems should identify, localize, and explain with structured evidence instead of collapsing everything into a single score.
The survey organizes AI-generated video into three broad paradigms. Each one changes which cues are most useful, and therefore which detection strategies remain reliable.
The evidence profile shifts as generation moves from real-carrier manipulation to fully synthetic video synthesis.
Real footage is preserved while localized regions or attributes are re-rendered. The strongest evidence is often spatially concentrated and residual.
Speech drives the visual stream, making synchrony, speaker consistency, and speech-conditioned motion more decisive than generic image artifacts.
No authentic carrier remains. Detection shifts toward long-range coherence, world plausibility, provenance, and language-guided verification.
The taxonomy is organized by the dominant evidence pathway used for decision-making, not by a rigid task label. It moves from low-level perception to high-level cognition while keeping the boundary between layers operational.
Visual view: Layers 1-2. Language view: Layers 3-4. The boundary depends on what evidence the detector actually needs.
Focuses on the visual modality and emphasizes statistical differences between AIGC-V and real videos. The pathway extends from frame-level intrinsic cues to spatiotemporal consistency across frames, forming the perceptual perspective for factual-fidelity verification.
Tests whether low-level visual signals still follow the statistical regularities of real videos, including frequency fingerprints, local geometry, physiological traces, and residual synthesis artifacts.
Boundary L1-2: Layer 1 focuses on frame-level intrinsic distributional cues.Models frame-sequence relations to test whether motion, behavior, and physical transitions unfold plausibly over time as a real recorded video.
Boundary L1-2 / L2-3: Layer 2 moves beyond frames to explicitly model spatiotemporal connections across the frame sequence, but it still stays within the visual modality and evaluates within-visual coherence.Defines a grounded verification pathway in which audio, speech, text, and visual evidence are first checked for within-video consistency, then extended to factual and knowledge-level verification when external evidence is required.
Performs within-video multimodal verification across speech, text, and visuals, from lip-speech synchrony to voice-face identity coherence and speech-text agreement.
Boundary L2-3 / L3-4: Layer 3 goes beyond vision and relies on within-video cross-modal agreement, judging whether different modalities are mutually consistent in describing the same semantics.Moves beyond within-video agreement to verify implied claims against external facts, commonsense, physical rules, and explicit evidence chains.
Boundary L3-4: Layer 4 further requires external facts, commonsense, or physics for reasoning-based verification, even when the modalities appear mutually consistent.The landscape figure highlights the transition from artifact matching in traditional deepfake detection to evidence-based semantic verification enabled by vision-language models and agentic pipelines.
These cards summarize the broader migration pattern across layers, while the table below adds a clearly marked partial snapshot through July 20, 2026.
| Year | L1 | L2 | L3 | L4 | Visual | Language | Language-View Share |
|---|---|---|---|---|---|---|---|
| 2020 | 5 | 7 | 1 | 0 | 12 | 1 | yearly / cumulative |
| 2021 | 5 | 6 | 4 | 0 | 11 | 4 | yearly / cumulative |
| 2022 | 4 | 3 | 2 | 0 | 7 | 2 | yearly / cumulative |
| 2023 | 3 | 3 | 4 | 0 | 6 | 4 | yearly / cumulative |
| 2024 | 9 | 5 | 4 | 3 | 14 | 7 | yearly / cumulative |
| 2025 | 1 | 22 | 20 | 15 | 23 | 35 | yearly / cumulative |
| 2026* | 12 | 32 | 30 | 9 | 44 | 39 | yearly / cumulative |
Complete publication years are shown through 2025. The 2026* row is an in-progress snapshot frozen on July 20, 2026; yearly and cumulative shares use the same video-first method-corpus counting rule as the paper.
Benchmark design should reveal why factual fidelity fails, where the failure occurs, and whether the system can ground its explanation with checkable evidence.
The deepest protocol tradition, emphasizing localized forensic residue, compression robustness, and cross-dataset transfer.
Smaller but tightly aligned with synchrony, speaker-content consistency, and timestamped localization.
Rapidly changing benchmarks targeting cross-generator transfer, semantic fabrication, plausibility, and world-level reasoning.
Adjacent diagnostics matter because trustworthy detection increasingly depends on physical plausibility and grounded explanations.
Browse the categorized paper and benchmark index supporting a survey corpus of 298 video-centered works, with verified venue-year labels and compact paper links.
Move across paradigms, detection layers, and benchmark families. April--July additions appear inside their canonical categories, and every new classification follows the detector's decisive evidence source.
Use Methods Overview for the taxonomy first, then switch to paradigms, layer pages, or benchmark families when you want a more specific slice of the literature.
Repository-level entry points and the complete flattened list.
Navigate by manipulation type and synthesis regime.
Follow the survey taxonomy from intrinsic cues to world-level reasoning.
Inspect evaluation families and adjacent diagnostic testbeds.
Select a route above to update the reading panel below.
Loading the selected library view.
A flat, paper-first index in the same style as the reference list. Detailed notes and extra metadata stay on the linked section pages.
arXiv 2025 FakeParts: a New Family of
AI-Generated DeepFakes. [paper]CIKM 2025 FakeChain: Exposing Shallow
Cues in Multi-Step Deepfake Detection. [paper]ICCV 2025 DynamicFace: High-quality
and consistent face swapping for image and video using composable 3D
facial priors. [paper]NeurIPS 2024 FuseAnyPart:
Diffusion-Driven Facial Parts Swapping via Multiple Reference Images.
[paper]arXiv 2025 SayAnything: Audio-Driven
Lip Synchronization with Conditional Video Diffusion. [paper]arXiv 2025 Audio-visual Controlled
Video Diffusion with Masked Selective State Spaces Modeling for Natural
Talking Head Generation. [paper]ACM MM 2025 Ditto: Motion-Space
Diffusion for Controllable Realtime Talking Head Synthesis. [paper]arXiv 2023 GeneFace: Generalized and
High-Fidelity Audio-Driven 3D Talking Face Synthesis. [paper]SIGGRAPH Asia 2022 VideoReTalking:
Audio-based Lip Synchronization for Talking Head Video Editing in the
Wild. [paper]Official release 2026 Official Launch
of Seedance 2.0. [paper]IJCV 2025 Show-1: Marrying pixel and
latent diffusion models for text-to-video generation. [paper]Official release 2025 Kling O1:
Unified Multimodal Video Model. [paper]Official release 2025 Veo 3. [paper]CVPR 2024 Grid diffusion models for
text-to-video generation. [paper]Official release 2024 Sora 2. [paper]Official release 2024 Introducing
Gen-3 Alpha: A New Frontier for Video Generation. [paper]Official release 2024 Dream Machine.
[paper]ICCV 2023 Scalable Diffusion Models
with Transformers. [paper]arXiv 2022 Video Diffusion Models. [paper]arXiv 2022 Imagen Video: High
Definition Video Generation with Diffusion Models. [paper]arXiv 2022 Make-A-Video: Text-to-Video
Generation without Text-Video Data. [paper]arXiv 2026 Lightweight
Complementary-Cue Fusion for Robust Video Face Forgery Detection. [paper]arXiv 2026 SpecSem-Net: Integrating
Spectral and Semantic Features for Robust AI-Generated Video Detection.
[paper]NeurIPS 2024 Freqblender: Enhancing
deepfake detection by blending frequency knowledge. [paper]ECCV 2024 Real appearance modeling for
more general deepfake detection. [paper]CVPR 2024 Beyond deepfake images:
Detecting ai-generated videos. [paper]AAAI 2023 Noise based deepfake
detection via multi-head relative-interaction. [paper]ECCV 2022 Hierarchical contrastive
inconsistency learning for deepfake video detection. [paper]CVPR 2021 MagDR: Mask-Guided Detection
and Reconstruction for Defending Deepfakes. [paper]CVPR 2021 Improving the efficiency and
robustness of deepfakes detection through precise geometric features.
[paper]ICASSP 2019 Exposing deep fakes using
inconsistent head poses. [paper]The Visual Computer 2024 Local
attention and long-distance interaction of rPPG for deepfake detection.
[paper]arXiv 2022 Visual Representations of
Physiological Signals for Fake Video Detection. [paper]arXiv 2021 Exposing deepfake with
pixel-wise ar and ppg correlation from faint signals. [paper]ICTC 2021 A study on effective use of
bpm information in deepfake detection. [paper]ACM MM 2020 Deeprhythm: Exposing
deepfakes with attentional visual heartbeat rhythms. [paper]arXiv 2020 Deepfakeson-phys: Deepfakes
detection based on heart rate estimation. [paper]IJCB 2020 How do the hearts of deep
fakes beat? Deep fake source detection via interpreting residuals with
biological signals. [paper]IEEE TPAMI 2020 Fakecatcher: Detection
of synthetic portrait videos using biological signals. [paper]ICCVW 2019 Predicting Heart Rate
Variations of Deepfake Videos using Neural ODE. [paper]WIFS 2018 In ictu oculi: Exposing ai
created fake videos by detecting eye blinking. [paper]arXiv 2026 G2VD: Generalizable
AI-Generated Video Detection via Counterfactual Intervention and Causal
Disentanglement. [paper]arXiv 2026 CUPID: Reconstructing UV
Texture Maps for Interpretable Person-of-Interest Deepfake Detection.
[paper]ACM IH&MMSec 2026 Seeing the
Unseen: Enhancing Synthetic Video Detector Transferability to New AI
Video Generation Models via Virtual Generators. [paper]CVPRW 2026 Forensic Rewiring:
Circuit-Level Shortcut Suppression for Generalizable Deepfake Detection.
[paper]CVPR 2026 DFD-HR: Generalizable
Deepfake Detection via Hierarchical Routing Learning. [paper]CVPR 2026 Beyond [CLS] Token:
Query-Driven Token-Level Forgery Purification for Generalizable Deepfake
Detection. [paper]CVPR 2026 Tutor-Student Reinforcement
Learning: A Dynamic Curriculum for Robust Deepfake Detection. [paper]ICML 2026 Real Data Lies: Unveiling
and Closing the Quality Shortcut in Generalizable AI-Generated Video
Detection. [paper]arXiv 2026 Video as Natural
Augmentation: Towards Unified AI-Generated Image and Video Detection.
[paper]ICDF2C 2025 Deepfake Forensics
Adapter: A Dual-Stream Network for Generalizable Deepfake Detection. [paper]WACV 2026 Deepfake Detection that
Generalizes Across Benchmarks. [paper]NeurIPS 2024 Can we leave deepfake
data behind in training deepfake detector? [paper]ECCV 2024 Fake It till You Make It:
Curricular Dynamic Forgery Augmentations Towards General Deepfake
Detection. [paper]CVPR 2024 Transcending Forgery
Specificity with Latent Space Augmentation for Generalizable Deepfake
Detection. [paper]CVPR 2024 Exploiting style latent
flows for generalizing deepfake video detection. [paper]arXiv 2024 Turns Out I'm Not Real:
Towards Robust Detection of AI-Generated Videos. [paper]ICCV 2023 Seeable: Soft discrepancies
and bounded contrastive learning for exposing deepfakes. [paper]ICCV 2023 Quality-agnostic deepfake
detection with intra-model collaborative learning. [paper]NeurIPS 2022 Ost: Improving
generalization of deepfake detection via one-shot test-time training.
[paper]CIKM 2020 Towards generalizable
deepfake detection with locality-aware autoencoder. [paper]CVPR 2026 SAGA: Source Attribution of
Generative AI Videos. [paper]CVPR 2026 Training-Free Detection of
Generated Videos via Spatial-Temporal Likelihoods. [paper]Expert Systems with Applications 2026
Robust Deepfake Detection in Compressed Videos with Scalable Network
Strategies. [paper]Scientific Reports 2026 A
Spatiotemporal Defect-Integrated Deepfake Video Detection and Forgery
Algorithm Attribution Model. [paper]IEEE TCSVT 2026 Robust AI-Generated
Video Detection via Color-Agnostic Contrastive Curriculum Learning with
Color-Aware Feature Modulation. [paper]Discover AI 2026 A Novel Temporal
Attention Regularization Loss for Robust and Explainable Deepfake
Detection. [paper]arXiv 2026 Rethinking the Readout:
Unlocking Video Backbones for AI-Generated Video Detection. [paper]ECCV 2026 (author-reported) SPLIT:
Training-Free AI-Generated and Partially Edited Video Detection via
Spatial Patch-Level Incoherence and Temporal Roughness. [paper]ICME 2026 (author-reported) LoCC:
Detection and Localization of Lip-Syncing Deepfakes via Counterfactual
Frame Consistency. [paper]CVPRW 2026 From Detector Evidence to
Language: Explainable Deepfake Video Detection. [paper]arXiv 2026 ReConFuse:
Reconstruction-Error Guided Semantic Fusion for AI-Generated Video
Detection. [paper]arXiv 2026 Scalable, Energy-Efficient
Optical-Neural Architecture for Multiplexed Deepfake Video Detection.
[paper]arXiv 2026 CAM-VFD: Cross-Attention
Multimodal Video Forgery Detection. [paper]arXiv 2026 Detecting AI-Generated
Videos with Spiking Neural Networks. [paper]arXiv 2026 DYMAPIA: A Multi-Domain
Framework for Detecting AI-Based Video Manipulation. [paper]ICASSP 2026 Towards Robust
Cross-Compression Deepfake Detection. [paper]ICLR 2026 Preserving Forgery
Artifacts: AI-Generated Video Detection at Native Scale. [paper]WACV 2026 SVD-Det: A Lightweight
Framework for Video Forgery Detection Using Semantic and Visual Defect
Cues. [paper]WACV 2026 Training-free Detection of
Text-to-video Generations via Over-coherence. [paper]Neurocomputing 2026 Detecting
AI-Generated Videos via Global Semantic Awareness and Inter-Frame
Semantic Consistency. [paper]Signal Processing: Image Communication 2026
3D Differential Decomposition for Video Deepfake Detection with Identity
Suppression. [paper]IEEE Access 2026 Beyond Frames:
3D-CoAtNet for Generalizable Deepfake Video Detection. [paper]IEEE TBIOM 2026 Dual Frequency-Guided
Spatiotemporal Feature Learning for Face Forgery Detection. [paper]Knowledge-Based Systems 2026 CAST:
Cross-Attentive Spatio-Temporal Feature Fusion for Deepfake Detection.
[paper] [arXiv]ICME 2025 GC-ConsFlow: Leveraging
Optical Flow Residuals and Global Context for Robust Deepfake Detection.
[paper]CVPR 2025 Generalizing deepfake video
detection with plug-and-play: Video-level blending and spatiotemporal
adapter tuning. [paper]arXiv 2025 Vulnerability-Aware
Spatio-Temporal Learning for Generalizable and Interpretable Deepfake
Video Detection. [paper]ECCV 2024 Learning natural consistency
representation for face forgery video detection. [paper]IJCV 2024 Learning spatiotemporal
inconsistency via thumbnail layout for face deepfake detection. [paper]ICME 2025 Decof: Generated video
detection via frame consistency. [paper]IEEE TIFS 2023 Dynamic difference
learning with spatio--temporal correlation for deepfake video detection.
[paper]ICCV 2023 Tall: Thumbnail layout for
deepfake video detection. [paper]IJCAI 2022 Region-Aware Temporal
Inconsistency Learning for DeepFake Video Detection. [paper]ICIAP 2022 Combining efficientnet and
vision transformers for video deepfake detection. [paper]AAAI 2022 Delving into the local:
Dynamic inconsistency learning for deepfake video detection. [paper]ACM MM 2021 Spatiotemporal
inconsistency learning for deepfake video detection. [paper]IJCAI 2021 Detecting Deepfake Videos
with Temporal Dropout 3DCNN. [paper]IJCAI 2021 Dynamic Inconsistency-aware
DeepFake Video Detection. [paper]WACV 2021 Interpretable and
trustworthy deepfake detection via dynamic prototypes. [paper]ECCV 2020 Two-branch recurrent network
for isolating deepfakes in videos. [paper]ACM MM 2020 Sharp multiple instance
learning for deepfake video detection. [paper]ICME 2020 Fsspotter: Spotting
face-swapped video by spatial and temporal clues. [paper]arXiv 2026 Revealing Artifacts via
Noise Amplification: A Novel Perspective for AI-Generated Video
Detection. [paper]ISWA 2026 VIDS-Guard: A Novel
Forensics-Aware Multi-Stream Transformer Framework for Robust Deepfake
Video Detection. [paper]arXiv 2026 Exposing and Mitigating
Temporal Attack in Deepfake Video Detection. [paper]ICML 2026 Beyond Pixels: Mining
Compressed Domain Artifacts for Efficient AI-Generated Video Detection.
[paper]arXiv 2026 Aletheia:
Physics-Conditioned Localized Artifact Attention (PhyLAA-X) for
End-to-End Generalizable and Robust Deepfake Video Detection. [paper]ICASSP 2026 Deep Video Frame
Interpolation Detection Via Event-Guided Temporal Analysis and
High-Frequency Artifacts. [paper]ECCV 2026 (author-reported) Moiré
Video Authentication: A Physical Signature Against AI Video Generation.
[paper]arXiv 2026 MPF-Net: Exposing
High-Fidelity AI-Generated Video Forgeries via Hierarchical Manifold
Deviation and Micro-Temporal Fluctuations. [paper]ICCV 2025 D3: Training-Free
AI-Generated Video Detection Using Second-Order Features. [paper]NeurIPS 2025 Physics-Driven
Spatiotemporal Modeling for AI-Generated Video Detection. [paper]NeurIPS 2025 AI-Generated Video
Detection via Perceptual Straightening. [paper]arXiv 2025 Beyond Spatial Frequency:
Pixel-wise Temporal Frequency-based Deepfake Video Detection. [paper]arXiv 2025 Leveraging Pre-Trained
Visual Models for AI-Generated Video Detection. [paper]arXiv 2025 De-Fake: Style based
Anomaly Deepfake Detection. [paper]arXiv 2025 Seeing What Matters:
Generalizable AI-generated Video Detection with Forensic-Oriented
Augmentation. [paper]arXiv 2025 VoD: Learning Volume of
Differences for Video-Based Deepfake Detection. [paper]WACV 2025 DiffFake: Exposing Deepfakes
using Differential Anomaly Detection. [paper]IEEE TMM 2024 DIP: diffusion learning
of inconsistency pattern for general deepfake detection. [paper]arXiv 2024 A quality-centric framework
for generic deepfake detection. [paper]CVPRW 2020 Towards untrusted social
video verification to combat deepfakes via face geometry consistency.
[paper]CVPR 2026 Learning Forgery-Aware Lip
Representations Without Forgery Priors. [paper]arXiv 2026 Interpretable Facial
Dynamics as Behavioral and Perceptual Traces of Deepfakes. [paper]arXiv 2026 BioLip:
Language-Generalizable Lip-Sync Deepfake Detection via Biomechanical
Constraint Violation Modeling. [paper]arXiv 2025 DeepFake Detection in
Dyadic Video Calls using Point of Gaze Tracking. [paper]WACVW 2026 Detecting Deepfake Talking
Heads from Facial Biometric Anomalies. [paper]arXiv 2025 When Deepfake Detection
Meets Graph Neural Network: a Unified and Lightweight Learning
Framework. [paper]CVPR 2025 Detecting Localized Deepfake
Manipulations Using Action Unit-Guided Video Representations. [paper]IEEE TCSVT 2023 Exploiting
complementary dynamic incoherence for deepfake video detection. [paper]CVPR 2021 Lips Don't Lie: A
Generalisable and Robust Approach To Face Forgery Detection. [paper]WIFS 2020 Detecting deep-fake videos
from appearance and behavior. [paper]arXiv 2020 Identity-driven deepfake
detection. [paper]arXiv 2020 Emotions Don't Lie: An
Audio-Visual Deepfake Detection Method Using Affective Cues. [paper]PRCV 2025 (LNCS 2026) Towards Highly
Generalized Lip-Sync Deepfake Detection via Detailed Audio-Visual
Inconsistency Analysis. [paper]Knowledge-Based Systems 2026 AMID:
Audio-Visual Deepfake Detection via Adaptive Multi-Dimensional
Interaction Modeling. [paper]IEEE TCSVT 2026 AVIR: Audio-Visual
Information Refinement for Robust Multimodal Deepfake Detection. [paper]arXiv 2026 Leave No Stone Unturned:
Uncovering Holistic Audio-Visual Intrinsic Coherence for Deepfake
Detection. [paper]IEEE TCSS 2026 VATS: Visual-Audio
Multitask Transformer With Specialty Audio Encoder for Multimodal
Deepfake Detection in CPSS. [paper]Neurocomputing 2026 AVSCNet: A
Dual-Branch Network for Synchronization Detection and Content
Consistency Learning in Audio-Video Forgery Detection. [paper]Neural Networks 2026 Emotion-Aware
Multimodal Deepfake Detection. [paper]arXiv 2026 Teacher--Student Structure
for Domain Adaptation in Ensemble Audio--Visual Video Deepfake
Detection. [paper]SAFE@CVPRW 2026 EMO-BOOST:
Emotion-Augmented Audio-Visual Features for Improved Generalization in
Deepfake Detection. [paper]ICML 2026 From Talking to Singing: A
New Challenge for Audio-Visual Deepfake Detection. [paper]arXiv 2026 Are DeepFakes Realistic
Enough? Exploring Semantic Mismatch as a Novel Challenge. [paper]arXiv 2026 Attribution-Guided
Multimodal Deepfake Detection via Cross-Modal Forensic Fingerprints. [paper]arXiv 2026 Listening Deepfake
Detection: A New Perspective Beyond Speaking-Centric Forgery Analysis.
[paper]arXiv 2026 Generalizing Video DeepFake
Detection by Self-Generated Audio-Visual Pseudo-Fakes. [paper]ICASSP 2026 MSCT: Differential
Cross-Modal Attention for Deepfake Detection. [paper] [arXiv]ICASSP 2026 AVATAR: Audio-Visual
Adaptive Fusion via Trained Agent Reinforcement for Multimodal Deepfake
Detection. [paper]ICASSP 2026 HybridMask: Facial-Guided
Cross-Modal Fusion for Multimodal Deepfake Detection. [paper]CVPR 2026 X-AVDT: Audio-Visual
Cross-Attention for Robust Deepfake Detection. [paper]ICASSP 2026 Leveraging Large
Multimodal Models for Audio-Video Deepfake Detection: A Pilot Study. [paper] [arXiv]Findings of EACL 2026 Revealing the
Truth with ConLLM for Detecting Multi-Modal Deepfakes. [paper]CVPR 2026 Investigating
Self-Supervised Representations for Audio-Visual Deepfake Detection. [paper]IEEE TIFS 2026 Toward Generalizable
Deepfake Detection via Forgery-Aware Audio-Visual Adaptation: A
Variational Bayesian Approach. [paper]Information Fusion 2026 Improving
Deepfake Detection with Predictive Inter-Modal Alignment and Feature
Reconstruction in Audio-Visual Asynchrony Scenarios. [paper]IEEE TIFS 2026 FauForensics: Boosting
Audio-Visual Deepfake Detection With Facial Action Units. [paper]arXiv 2025 PIA: Deepfake Detection
Using Phoneme-Temporal and Identity-Dynamic Analysis. [paper]ACM MM 2025 KLASSify to Verify:
Audio-Visual Deepfake Detection Using SSL-based Audio and Handcrafted
Visual Features. [paper]arXiv 2025 CAD: A General Multimodal
Framework for Video Deepfake Detection via Cross-Modal Alignment and
Distillation. [paper]AAAI 2025 Multi-modal deepfake
detection via multi-task audio-visual prompt learning. [paper]CVPRW 2024 Lost in Translation:
Lip-Sync Deepfake Detection from Audio-Video Mismatch. [paper]CVPR 2024 AVFF: Audio-Visual Feature
Fusion for Video Deepfake Detection. [paper]Interspeech 2024 Zero-Shot Fake Video
Detection by Audio-Visual Consistency. [paper]ACM TOMM 2023 Voice-Face Homogeneity
Tells Deepfake. [paper]IEEE URTC 2023 Integrating
Audio-Visual Features for Multimodal Deepfake Detection. [paper]APSIPA ASC 2022 Lip Sync Matters: A
Novel Multimodal Forgery Detector. [paper]CVPRW 2023 Audio-Visual
Person-of-Interest DeepFake Detection. [paper]arXiv 2022 An Audio-Visual Attention
Based Multimodal Network for Fake Talking Face Videos Detection. [paper]ICCV 2021 Joint Audio-Visual Deepfake
Detection. [paper]ICME 2021 DeepFake Videos Detection
Using Self-Supervised Decoupling Network. [paper]CVPRW 2021 Detecting Deep-Fake Videos
From Aural and Oral Dynamics. [paper]IEEE TIFS 2021 Preventing DeepFake
Attacks on Speaker Authentication by Dynamic Lip Movement Analysis. [paper]ACM MM 2020 Not made for each other-
Audio-Visual Dissonance-based Deepfake Detection and Localization. [paper]Future Internet 2026 Video Deepfake
Detection Based on Multimodality Semantic Consistency Fusion. [paper]arXiv 2026 CMTA: Leveraging
Cross-Modal Temporal Artifacts for Generalizable AI-Generated Video
Detection. [paper]arXiv 2026 ATSS: Detecting
AI-Generated Videos via Anomalous Temporal Self-Similarity. [paper]arXiv 2025 T^3SVFND: Towards an
Evolving Fake News Detector for Emergencies with Test-time Training on
Short Video Platforms. [paper]CVPR 2025 Unleashing the Potential of
Consistency Learning for Detecting and Grounding Multi-Modal Media
Manipulation. [paper]arXiv 2025 Consistency-aware Fake
Videos Detection on Short Video Platforms. [paper]CVPR 2026 GEM-TFL: Bridging Weak and
Full Supervision for Forgery Localization through EM-Guided
Decomposition and Temporal Refinement. [paper]Alexandria Engineering Journal 2026
CCFormer: A Cascaded Transformer Framework for Precise Temporal
Audio-Visual Deepfake Localization. [paper]arXiv 2026 EVAS: Efficient Multimodal
Temporal Forgery Localization via Audio-Visual Synergy and Steered
Boundary Calibration. [paper]ECCV 2026 (author-reported) MG-RWKV:
Multi-Grained Context-Aware RWKV for Temporal Forgery Localization. [paper]CVPR 2026 Inconsistency-Aware
Multimodal Schrödinger Bridge for Deepfake Localization. [paper]arXiv 2026 Towards Multi-Modal Forgery
Representation Learning for AI-Generated Video Detection and
Localization. [paper]arXiv 2026 Divide and Conquer:
Multimodal Video Deepfake Detection via Cross-Modal Fusion and
Localization. [paper]WACV 2026 A-V Representation Learning
via Audio Shift Prediction for Multimodal Deepfake Detection and
Temporal Localization. [paper]WACV 2026 AuViRe: Audio-visual Speech
Representation Reconstruction for Deepfake Temporal Localization. [paper]ACM MM 2025 HOLA: Enhancing
Audio-visual Deepfake Detection via Hierarchical Contextual Aggregations
and Efficient Pre-training. [paper]ACM MM 2025 A Multimodal Deviation
Perceiving Framework for Weakly-Supervised Temporal Forgery
Localization. [paper]arXiv 2025 SpeechForensics:
Audio-Visual Speech Representation Learning for Face Forgery Detection.
[paper]arXiv 2025 Weakly Supervised
Multimodal Temporal Forgery Localization via Multitask Learning. [paper]CVPR 2025 Circumventing Shortcuts in
Audio-visual Deepfake Detection Datasets with Unsupervised Learning. [paper]ICASSP 2025 Audio-Visual Deepfake
Detection With Local Temporal Inconsistencies. [paper]arXiv 2024 DiMoDif: Discourse
Modality-information Differentiation for Audio-visual Deepfake Detection
and Localization. [paper]ICASSP 2024 Cross-Modality and
Within-Modality Regularization for Audio-Visual DeepFake Detection. [paper]CVPR 2023 Self-supervised video
forensics by audio-visual anomaly detection. [paper]International Journal of Digital Crime and Forensics 2026
Vision Forgery Trace Enhanced VLMs for Generalized AIGC Video Detection.
[paper]CVPR 2026 Unleashing Vision-Language
Semantics for Deepfake Video Detection. [paper]ICML 2025 Unlocking the Capabilities
of Large Vision-Language Models for Generalizable and Explainable
Deepfake Detection. [paper]WACV 2026 AuthGuard: Generalizable
Deepfake Detection via Language Guidance. [paper]AAAI 2025 Standing on the Shoulders of
Giants: Reprogramming Visual-Language Model for General Deepfake
Detection. [paper]IJCV 2025 DeepFake-Adapter: Dual-Level
Adapter for DeepFake Detection. [paper]BMVC 2024 Prompt-guided Multi-modal
contrastive learning for Cross-compression-rate Deepfake Detection. [paper]CVPRW 2024 Can ChatGPT Detect
DeepFakes? A Study of Using Multimodal Large Language Models for Media
Forensics. [paper]APSIPA TSIP 2025 How Good is ChatGPT
at Audiovisual Deepfake Detection: A Comparative Study of ChatGPT, AI
Models and Human Perception. [paper]ECCV 2026 (author-reported) SafeGuard:
A Multi-Agent Perception-Reasoning Framework for Social-Risk
AI-Generated Video Detection. [paper]CVPR 2026 VMD-FACT: A New Video
Dataset and MLLM-Based Method for Detecting Realistic AI-Generated Video
Misinformation. [paper]ICML 2026 Explainable Forensics of
Manipulated Segments in Untrimmed Long Videos. [paper]ICML 2026 Hermes: An Evidence-Driven
Agentic Framework for Trustworthy and Explainable AI-Generated Video
Detection. [paper]CVPR 2026 CoCoVideo: The High-Quality
Commercial-Model-Based Contrastive Benchmark for AI-Generated Video
Detection. [paper]arXiv 2026 DVAR: Adversarial
Multi-Agent Debate for Video Authenticity Detection. [paper]arXiv 2025 DeepAgent: A Dual Stream
Multi Agent Fusion for Robust Multimodal Deepfake Detection. [paper]arXiv 2025 Memory-Anchored Multimodal
Reasoning for Explainable Video Forensics. [paper]arXiv 2025 DAVID-XR1: Detecting
AI-Generated Videos with Explainable Reasoning. [paper]arXiv 2025 LAVID: An Agentic LVLM
Framework for Diffusion-Generated Video Detection. [paper]ICML 2026 VideoVeritas: AI-Generated
Video Detection via Perception Pretext Reinforcement Learning. [paper]CVPR 2026 Skyra: AI-Generated Video
Detection via Grounded Artifact Reasoning. [paper]arXiv 2025 VidGuard-R1: AI-Generated
Video Detection and Explanation via Reasoning MLLMs and RL. [paper]arXiv 2025 EDVD-LLaMA: Explainable
Deepfake Video Detection via Multimodal Large Language Model Reasoning.
[paper]arXiv 2025 Learning Human-Perceived
Fakeness in AI-Generated Videos via Multimodal LLMs. [paper]arXiv 2025 Veritas: Generalizable
Deepfake Detection via Pattern-Aware Reasoning. [paper]arXiv 2025 BusterX++: Towards Unified
Cross-Modal AI-Generated Content Detection and Explanation with MLLM.
[paper]arXiv 2025 BusterX: MLLM-Powered
AI-Generated Video Forgery Detection and Explanation. [paper]arXiv 2024 X2-DFD: A framework for
eXplainable and eXtendable Deepfake Detection. [paper]arXiv 2026 HumanForge: A Human-Centric
Deepfake Video Benchmark with Multi-Agent Forgery Rationales. [paper]CVPR 2026 DeepfakeImpact: A Two-Stage
Benchmark with Real-World Impact in Deepfake Detection. [paper]arXiv 2026 Detecting Temporally
Localized Manipulations in Authentic Video Streams. [paper]ICML 2026 Explainable Forensics of
Manipulated Segments in Untrimmed Long Videos. [paper]CVPR 2026 ActivityForensics: A
Comprehensive Benchmark for Localizing Manipulated Activity in Videos.
[paper]Machine Vision and Applications 2026
Cross-Dataset Video Deepfake Detection Using Transformer and CNN
Architectures (introduces ReenactFaces). [paper]CVPR Findings 2026 Beyond Static
Artifacts: A Forensic Benchmark for Video Deepfake Reasoning in Vision
Language Models. [paper]WACV 2026 Exddv: A new dataset for
explainable deepfake detection in video. [paper]ECCV 2024 Common sense reasoning for
deepfake detection. [paper]CVPR 2025 Ai-face: A million-scale
demographically annotated ai-generated face dataset and fairness
benchmark. [paper]NeurIPS 2023 DeepfakeBench: A
Comprehensive Benchmark of Deepfake Detection. [paper]CVPR 2023 DF-Platter: Multi-Face
Heterogeneous Deepfake Dataset. [paper]WACV 2023 A Continual Deepfake
Detection Benchmark: Dataset, Methods, and Essentials. [paper]ICCV 2021 Kodf: A large-scale korean
deepfake detection dataset. [paper]CVPR 2021 ForgeryNet: A Versatile
Benchmark for Comprehensive Forgery Analysis. [paper]ACM MM 2020 WildDeepfake: A
Challenging Real-World Dataset for Deepfake Detection. [paper]arXiv 2020 The DeepFake Detection
Challenge (DFDC) Dataset. [paper]CVPR 2020 DeeperForensics-1.0: A
Large-Scale Dataset for Real-World Face Forgery Detection. [paper]CVPR 2020 Celeb-df: A large-scale
challenging dataset for deepfake forensics. [paper]ICCV 2019 Faceforensics++: Learning to
detect manipulated facial images. [paper]CVPR 2026 VMD-FACT: A New Video
Dataset and MLLM-Based Method for Detecting Realistic AI-Generated Video
Misinformation. [paper]CVPR 2026 AVFakeBench: A Comprehensive
Audio-Video Forgery Detection Benchmark for AV-LMMs. [paper]CVPR 2026 X-AVDT: Audio-Visual
Cross-Attention for Robust Deepfake Detection. [paper]ACM MM 2025 Av-deepfake1m++: A
large-scale audio-visual deepfake benchmark with real-world
perturbations. [paper]Interspeech 2025 VCapAV: A
Video-Caption Based Audio-Visual Deepfake Detection Dataset. [paper]arXiv 2025 Memory-Anchored Multimodal
Reasoning for Explainable Video Forensics. [paper]ACM MMAD Workshop 2025 SocialDF:
Benchmark Dataset and Detection Model for Mitigating Harmful Deepfake
Content on Social Media Platforms. [paper]arXiv 2025 Tell me Habibi, is it Real
or Fake? [paper]arXiv 2025 MAVOS-DD: Multilingual
Audio-Video Open-Set Deepfake Detection Benchmark. [paper]arXiv 2025 Beyond Face Swapping: A
Diffusion-Based Digital Human Benchmark for Multimodal Deepfake
Detection. [paper]WACV 2026 TalkingHeadBench: A
Multi-Modal Benchmark & Analysis of Talking-Head DeepFake Detection.
[paper]ACM MM 2024 AV-Deepfake1M: A
large-scale LLM-driven audio-visual deepfake dataset. [paper]arXiv 2024 WWW: Where, Which and
Whatever Enhancing Interpretability in Multimodal Deepfake Detection.
[paper]DICTA 2022 Do you really mean that?
content driven audio-visual deepfake dataset and multimodal method for
temporal forgery localization. [paper]arXiv 2021 FakeAVCeleb: A Novel
Audio-Video Multimodal Deepfake Dataset. [paper]CVPR 2026 Training-Free Detection of
Generated Videos via Spatial-Temporal Likelihoods. [paper]ECCV 2026 (author-reported) SafeGuard:
A Multi-Agent Perception-Reasoning Framework for Social-Risk
AI-Generated Video Detection. [paper]arXiv 2026 Auditing Generalization in
AI-Generated Video Detection: A Six-Control Protocol and the VidAudit
Toolkit. [paper]arXiv 2026 Revealing Artifacts via
Noise Amplification: A Novel Perspective for AI-Generated Video
Detection. [paper]CVPR 2026 FVBench: Benchmarking
Deepfake Video Detection Capability of Large Multimodal Models. [paper]arXiv 2026 Artifact-Bench: Evaluating
MLLMs on Detecting and Assessing the Artifacts of AI-Generated Videos.
[paper]CVPR 2026 CoCoVideo: The High-Quality
Commercial-Model-Based Contrastive Benchmark for AI-Generated Video
Detection. [paper]arXiv 2026 Advancing Reliable
Synthetic Video Detection: Insights from the SAFE Challenge. [paper]ICLR 2026 Preserving Forgery
Artifacts: AI-Generated Video Detection at Native Scale. [paper]arXiv 2026 SynthForensics: A
Multi-Generator Benchmark for Detecting Synthetic Video Deepfakes. [paper]ICML 2026 VideoVeritas: AI-Generated
Video Detection via Perception Pretext Reinforcement Learning. [paper]CVPR 2026 Your One-Stop Solution for
AI-Generated Video Detection. [paper]CVPR 2026 Skyra: AI-Generated Video
Detection via Grounded Artifact Reasoning. [paper]arXiv 2025 Video Reality Test: Can
AI-Generated ASMR Videos fool VLMs and Humans? [paper]arXiv 2025 EDVD-LLaMA: Explainable
Deepfake Video Detection via Multimodal Large Language Model Reasoning.
[paper]ACM MM 2025 AEGIS: Authenticity
Evaluation Benchmark for AI-Generated Video Sequences. [paper]arXiv 2025 Learning Human-Perceived
Fakeness in AI-Generated Videos via Multimodal LLMs. [paper]arXiv 2025 BusterX++: Towards Unified
Cross-Modal AI-Generated Content Detection and Explanation with MLLM.
[paper]arXiv 2025 GenWorld: Towards Detecting
AI-generated Real-world Simulation Videos. [paper]arXiv 2025 Ivy-fake: A unified
explainable framework and benchmark for image and video aigc detection.
[paper]arXiv 2025 DAVID-XR1: Detecting
AI-Generated Videos with Explainable Reasoning. [paper]arXiv 2025 BusterX: MLLM-Powered
AI-Generated Video Forgery Detection and Explanation. [paper]arXiv 2025 Deepfake-eval-2024: A
multi-modal in-the-wild benchmark of deepfakes circulated in 2024. [paper]AAAI 2026 Genvidbench: A challenging
benchmark for detecting ai-generated video. [paper]NeurIPS 2024 On Learning Multi-Modal
Forgery Representation for Diffusion Generated Video Detection. [paper]arXiv 2024 Distinguish any fake
videos: Unleashing the power of large-scale data and motion features.
[paper]arXiv 2024 DeMamba: AI-Generated Video
Detection on Million-Scale GenVideo Benchmark. [paper]arXiv 2024 Detecting AI-Generated
Video via Frame Consistency. [paper]arXiv 2026 Physion-Eval: Evaluating
Physical Realism in Generated Video via Human Reasoning. [paper]ICLR 2026 VideoPhy-2: A Challenging
Action-Centric Physical Commonsense Evaluation in Video Generation. [paper]arXiv 2025 PhyWorldBench: A Physical
Realism Benchmark for Text-to-Video Generation. [paper]arXiv 2025 T2VPhysBench: A
First-Principles Benchmark for Physical Consistency in Text-to-Video
Generation. [paper]arXiv 2025 Morpheus: Benchmarking
Physical Reasoning of Video Generative Models with Real Physical
Experiments. [paper]ICML 2025 Impossible Videos. [paper]arXiv 2025 Do generative video models
understand physical principles? [paper]ICLR 2025 VideoPhy: Evaluating
Physical Commonsense for Video Generation. [paper]arXiv 2025 SVBench: Evaluation of
Video Generation Models on Social Reasoning. [paper]arXiv 2025 VideoVerse: How Far is Your
T2V Generator from a World Model? [paper]arXiv 2025 T2VWorldBench: A Benchmark
for Evaluating World Knowledge in Text-to-Video Generation. [paper]arXiv 2024 Is Your World Simulator a
Good Story Presenter? A Consecutive Events-Based Benchmark for Future
Long Video Generation. [paper]ICML 2025 WorldSimBench: Towards Video
Generation Models as World Simulators. [paper]ICML 2025 Towards World Simulator:
Crafting Physical Commonsense-Based Benchmark for Video Generation. [paper]NeurIPS 2025 VideoHallu: Evaluating
and Mitigating Multi-modal Hallucinations on Synthetic Video
Understanding. [paper]arXiv 2025 PhyDetEx: A Benchmark
Dataset and Method for Detecting and Explaining Physical Plausibility in
Text-to-Video Models. [paper]arXiv 2025 SPOTLIGHT: Identifying and
Localizing Video Generation Errors Using VLMs. [paper]arXiv 2025 TRAVL: A Recipe for Making
Video-Language Models Better Judges of Physics Implausibility. [paper]arXiv 2026 Continuously Evolving
Deepfake Detection: An Architecture and Public-Benchmark Evaluation of a
Dynamic Detection System. [paper]arXiv 2026 Dataset Biases and Shortcut
Learning in Motion-Based AI-Generated Video Detection. [paper]arXiv 2026 When AUC Misleads:
Polarization-Aware Evaluation of Deepfake Detectors under Domain Shift.
[paper]arXiv 2026 Auditing Generalization in
AI-Generated Video Detection: A Six-Control Protocol and the VidAudit
Toolkit. [paper]arXiv 2026 GIFGuard: Proactive
Forensics against Deepfakes in Facial GIFs via Spatiotemporal
Watermarking. [paper]arXiv 2026 LAVA: Layered Audio-Visual
Anti-tampering Watermarking for Robust Deepfake Detection and
Localization. [paper]ROMCIR 2026 Beyond Seeing Is
Believing: On Crowdsourced Detection of Audiovisual Deepfakes. [paper]The next phase is not just stronger clip-level discrimination. It requires diagnostic evaluation, claim-level supervision, unified explainable detection, and deployment protocols that preserve calibration when evidence is weak, conflicting, or incomplete.
Move beyond clip-level AUC or EER. Evaluation should reveal which factual proposition fails, where the violation occurs, and how brittle the detector becomes under shift, transfer, codec changes, and modern synthesis pipelines.
Benchmarks should move from clip labels to checkable claims with timestamped evidence, targeted stress tests, and continuously refreshed generator pools rather than static one-off test sets.
Trustworthy AIGC-V detection is a two-pathway problem: perceptual evidence plus fact-level verification. Cross-layer fusion should preserve low-level traces, localized mismatches, and tested claims as individually inspectable evidence objects.
Systems should identify, localize, and explain with calibrated uncertainty, provenance-aware cross-checks, and abstention when evidence remains incomplete or internally contradictory.
We welcome missing papers, corrected venue records, code links, and benchmark updates. Please identify the proposed taxonomy layer and the evidence actually consumed by the detector.
The paper should centrally address AI-generated or manipulated video detection, localization, explanation, or detector evaluation. Audio-visual work qualifies when a video clip is genuinely analyzed.
Image-first deepfake papers and broad synthetic-media benchmarks are excluded when video is only a small auxiliary modality. Venue and first-public dates should resolve to primary sources.
If you use this survey or its taxonomy in your work, please cite it as follows.
@inproceedings{Hou_2026,
title={Detecting AI-Generated Video: A Vision–Language Dual-View Survey},
url={http://dx.doi.org/10.18653/v1/2026.findings-acl.1613},
DOI={10.18653/v1/2026.findings-acl.1613},
booktitle={Findings of the Association for Computational Linguistics: ACL 2026},
publisher={Association for Computational Linguistics},
author={Hou, Dylan Xinming and Zhang, Juntian and Gu, Xu and Wu, Yichen and Lukas, Nils and Xia, Gus and Chen, Xiuying and Liu, Yuhan},
year={2026},
pages={32221–32255} }