A Review of Tools and Technologies to Combat Deepfakes
Open accessThe DOI matches its Crossref record; 41 of the 41 references it cites were found in a registry or an index; 56 of the 115 sources its claims cite were checked for support.
Abstract
Reviews passive detection, active provenance, watermarking, cryptographic manifests, robustness, datasets, benchmarks, deployment and policy constraints, concluding layered defenses are needed.
Overview
CoChat Summary
TL;DR
This review argues that no single deepfake countermeasure is sufficient in realistic adversarial settings; the strongest practical strategy is a layered defense combining provenance, watermarking, content-based detection, and human oversight. It synthesizes evidence across modalities, benchmarks, standards, and policy, and emphasizes that detector generalization remains fragile under cross-dataset shift, unseen generators, compression, and adversarial post-processing.
Key findings
- The review’s four main findings are: content-based detectors can perform strongly in-domain but lose reliability under cross-dataset shift, unseen generators, compression, and adversarial post-processing; provenance/authentication gives the strongest affirmative evidence when signatures and trust chains validate; watermarking supports scalable transparency but is vulnerable to removal, spoofing, and fragmentation; and the most robust practical strategy is a layered architecture combining provenance, watermark checks, content-based analysis, uncertainty estimation, logging, and human escalation.
- Deepfake detection benchmarks show a large generalization gap: for example, Table 1 reports within-domain vs. cross-domain Avg AUC of 0.9450 vs. 0.7718 for Xception, 0.9449 vs. 0.7645 for F3-Net, 0.9391 vs. 0.6985 for Face X-ray, and 0.6097 vs. 0.6551 for Meso4.
- Dataset and benchmark evidence emphasizes scale and realism: DFDC is described as having over 100,000 clips from 3,426 paid actors, DeeperForensics-1.0 as 60,000 videos / 17.6M frames, ForgeryNet as 2.9M images and 221,247 videos, and WildDeepfake as 707 deepfake videos and 7,314 face sequences.
- The paper reports deployment-relevant tradeoffs: CNN-based detectors at 224 × 224 can reach about 88–90% accuracy at roughly 28–30 fps, while transformer-based, multimodal, and hybrid detectors reach approximately 92–95% accuracy but often only 18–22 fps.
Methods
- Narrative literature review with an evidence-oriented approach.
- Sources were gathered via structured searches of Google Scholar, arXiv, IEEE Xplore, ACM Digital Library, CVF Open Access, OpenReview, and official repositories/portals from standards bodies and public institutions; emphasis was on 2018–2026.
- Included peer-reviewed papers, benchmark papers, standards/specifications, official reports, dataset documentation, and open-source tooling; excluded speculative, marketing-only, duplicated, or unverifiable claims.
Limitations
- The paper states that parts of the most operationally important ecosystem—platform enforcement, large-scale detector performance, and real-time deployment constraints—are under-documented in peer-reviewed venues, forcing reliance on official documentation and selected technical reports.
- It states that performance values across papers are often not directly comparable because preprocessing, face detection, sampling strategies, and frame-level vs. video-level aggregation differ.
- It states that policy interpretation is time sensitive because EU legal texts and Commission guidance can evolve beyond the cited snapshots.
- Evident from the text: this is a narrative review rather than a new experimental study, and it reports “No data used in the study,” so its conclusions depend on the quality and comparability of the underlying literature.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
53 Valid · 2 Not valid · 56 Not checked · 41 Own results and interpretations
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“As detectors matured, research increasingly separated approaches into naïve end-to-end classifiers versus methods that explicitly model forgery artifacts such as blending boundaries, warping, or inconsistent face boundaries [8].”
What the sources say
- Partly supportedcites [8] Yan, Z., Zhang, Y., Yuan, X., Lyu, S., and Wu, B. (2023). Deepfakebench: A comprehensive benchmark of deepfake detection. arXiv. sourceExperimental — accuracy not yet measured
Related Work (p. 2) distinguishes naïve CNN classifiers from spatial methods, including methods focused on blending artifacts. The algorithm descriptions (p. 18) also identify a detector targeting face-warping artifacts. The paper uses a three-category taxonomy that additionally includes frequency detectors, and the passages do not specifically identify inconsistent face boundaries as an artifact modeled by these methods. (LLM, conf=0.83)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Attackers can cheaply generate high volumes of plausible false media, while defenders must either prove authenticity or detect manipulation under uncertainty, including adversarial post-processing [2].”
What the sources say
- Partly supportedcites [2] Verdoliva, 2020, Media forensics and deepfakes: An overview, IEEE J. Sel. Top. Signal Process. sourceExperimental — accuracy not yet measured
The abstract (p. 1) says freely available software lets people without special skills create realistic fake media, and the Introduction (p. 1) says verifying integrity can be difficult. Counterforensics (pp. 15–16) describes attacks that mislead detectors. The paper does not establish that attackers can generate false media in high volumes; it also qualifies ease of deepfake creation by the need for large amounts of data. (LLM, conf=0.86)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“This asymmetry is compounded by the open-world setting of online distribution, where content is frequently resized, recompressed, transcoded, screenshotted, or otherwise transformed in ways that can erase fragile forensic signals and complicate model assumptions [3].”
What the sources say
- Partly supportedcites [3] National Institute of Standards and Technology (2024). Reducing Risks Posed by Synthetic Content: An Overview of Technical Approaches to Digital Content Transparency (NIST AI 100-4). sourceExperimental — accuracy not yet measured
Split into 2 atoms — 0 confirmed, 2 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“A recurring result across surveys, benchmarks, and institutional guidance is that detector performance often drops substantially in cross-dataset or in-the-wild conditions, particularly under dataset shift, novel generators, or post-processing [8].”
What the sources say
- Supportedcites [8] Yan, Z., Zhang, Y., Yuan, X., Lyu, S., and Wu, B. (2023). Deepfakebench: A comprehensive benchmark of deepfake detection. arXiv. source
Split into 2 atoms — 2 confirmed, 0 partially supported, 0 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.
plainstates a possible cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Contemporary research and practitioner consensus increasingly treats deepfake defense as an end-to-end socio-technical system problem, rather than a narrow classification problem.”
What the sources say
- Partly supportedcites [3] National Institute of Standards and Technology (2024). Reducing Risks Posed by Synthetic Content: An Overview of Technical Approaches to Digital Content Transparency (NIST AI 100-4). sourceExperimental — accuracy not yet measured
The Conclusion (section 7; page not provided) says that technical approaches alone are not comprehensive and must be paired with social efforts and coordination across the content value chain. Section 6 also places transparency measures across AI lifecycle stages. These passages support the socio-technical, end-to-end framing, but do not establish an increasing research and practitioner consensus. (LLM, conf=0.84)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
53 more claims · 21 with their cited sources checked
Create a free account to see all 152 claims and to verify the 32 references still open.
- Probably not a substantial claim. 57
- It may state two separate claims as one. 63
- The quote only partly states this claim. 1
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ 8. Dataset and Benchmark Landscape
“| DeepSpeak | 2204 | Audio + video | Consented webcam-style audiovisual deepfakes with identity matching”
§ 14. Conclusions
“43. Barrington, S.; Bohacek, M.; Farid, H. The DeepSpeak Dataset. arXiv 2024 , arXiv:2408.05366. [CrossRef]”
Citations
Cites
FaceForensics++: Learning to Detect Manipulated Facial Images
Rossler, A., Cozzolino, D., Verdoliva, L., Riess, C., Thies, J., & Niessner, M. (2019) · 2019 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2019.00009 · Cited by 3,314 · cited 1×
“FaceForensics++ introduced a widely used benchmark for facial manipulation detection, providing a standardized evaluation setting across multiple manipulation methods and compression levels [34].”
MesoNet: a Compact Facial Video Forgery Detection Network
Afchar, D., Nozick, V., Yamagishi, J., & Echizen, I. (2018) · 2018 IEEE International Workshop on Information Forensics and Security (WIFS) · doi:10.1109/wifs.2018.8630761 · Cited by 1,872 · cited 1×
“Compact architectures such as MesoNet were designed to capture mesoscopic properties and report strong detection rates on then-common deepfake styles, illustrating the viability of lightweight inference for video tampering detection under compression constraints [7].”
Celeb-DF: A Large-Scale Challenging Dataset for DeepFake Forensics
Li, Y., Yang, X., Sun, P., Qi, H., & Lyu, S. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.00327 · Cited by 1,846 · cited 1×
“Celeb-DF aimed to increase realism and difficulty by producing higher-quality deepfakes aligned with internet-circulated content characteristics [36].”
Thinking in Frequency: Face Forgery Detection by Mining Frequency-Aware Clues
Qian, Y., Yin, G., Sheng, L., Chen, Z., & Shao, J. (2020) · Lecture Notes in Computer Science · doi:10.1007/978-3-030-58610-2_6 · Cited by 1,058 · cited 1×
“F3-Net (Frequency in Face Forgery Network) exemplifies this family by mining frequency-aware clues and is incorporated into standardized benchmarking as a representative frequency detector [9].”
Media Forensics and DeepFakes: An Overview
Verdoliva, L. (2020) · IEEE Journal of Selected Topics in Signal Processing · doi:10.1109/jstsp.2020.3002101 · Cited by 979 · cited 2×
“Survey work highlights that deepfakes accelerate traditional forensics vs. anti-forensics dynamic, shifting the problem from detecting relatively constrained edits to detecting high-capacity generative processes that can synthesize or rewrite content with semantically coherent detail [2].”
DeeperForensics-1.0: A Large-Scale Dataset for Real-World Face Forgery Detection
Jiang, L., Li, R., Wu, W., Qian, C., & Loy, C. C. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.00296 · Cited by 599 · cited 1×
“0 emphasized scale and real-world perturbations, supporting robustness evaluation via large video volume and diverse degradations [37].”
FakeCatcher: Detection of Synthetic Portrait Videos using Biological Signals
Ciftci, U. A., Demir, I., & Yin, L. (2024) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2020.3009287 · Cited by 499
WildDeepfake: A Challenging Real-World Dataset for Deepfake Detection
Zi, B., Chang, M., Chen, J., Ma, X., & Jiang, Y.-G. (2020) · Proceedings of the 28th ACM International Conference on Multimedia · doi:10.1145/3394171.3413769 · Cited by 448
Towards Universal Fake Image Detectors that Generalize Across Generative Models
Ojha, U., Li, Y., & Lee, Y. J. (2023) · 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52729.2023.02345 · Cited by 428 · cited 1×
“Complementary research argues for universal fake-image detectors that generalize across generative model families using feature spaces not explicitly trained for real-vs-fake discrimination [41].”
The Deepfake Detection Challenge (DFDC) Preview Dataset
Dolhansky, B., Howes, R., Pflaum, B., Baram, N., & Ferrer, C. C. (2019) · arXiv · doi:10.48550/arxiv.1910.08854 · Cited by 413
AASIST: Audio Anti-Spoofing Using Integrated Spectro-Temporal Graph Attention Networks
Jung, J.-W., Heo, H.-S., Tak, H., Shim, H.-J., Chung, J. S., Lee, B.-J., Yu, H.-J., & Evans, N. (2022) · ICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) · doi:10.1109/icassp43922.2022.9747766 · Cited by 399 · cited 1×
“Model architectures such as AASIST (Audio Anti-Spoofing using Integrated Spectro-Temporal Graph Attention Networks) aim to jointly exploit spectral and temporal artifacts via graph attention mechanisms, explicitly motivating single-system alternatives to ensembles [13].”
Towards World Simulator: Crafting Physical Commonsense-Based Benchmark for Video GenerationPreprint
Meng, F., Liao, J., Tan, X., Shao, W., Lu, Q., Zhang, K., Cheng, Y., Li, D., Yu, Q., & Luo, P. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2410.05363 · Cited by 180
ForgeryNet: A Versatile Benchmark for Comprehensive Forgery Analysis
He, Y., Gan, B., Chen, S., Zhou, Y., Yin, G., Song, L., Sheng, L., Shao, J., & Liu, Z. (2021) · 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr46437.2021.00434 · Cited by 161 · cited 1×
“ForgeryNet pushed scale further, reporting millions of images and hundreds of thousands of videos across multiple tasks and manipulation types [38].”
Deepfake: definitions, performance metrics and standards, datasets, and a meta-review
Altuncu, E., Franqueira, V. N. L., & Li, S. (2024) · Frontiers in Big Data · doi:10.3389/fdata.2024.1400024 · Cited by 134 · cited 2×
“The term deepfake has evolved from describing face-swapped videos to serving as an umbrella for AI-generated or AI-manipulated media, including images, video, audio voice cloning, and multimodal composites [1].”
ASVspoof 2021: Automatic Speaker Verification Spoofing and Countermeasures Challenge Evaluation Plan
Delgado, H., Evans, N., Kinnunen, T., Lee, K. A., Liu, X., Nautsch, A., Patino, J., Sahidullah, M., Todisco, M., Wang, X., et al. (2021) · arXiv · doi:10.48550/arxiv.2109.00535 · Cited by 129 · cited 3×
“ASVspoof 2021 specifies min t-DCF for ASV-constrained tasks (LA/PA) and EER for the deepfake speech task (DF), making explicit that different operational settings require different objective functions [12].”
InfluentialGenimage: A million-scale benchmark for detecting ai-generated image
Zhu, M., Chen, H., Yan, Q., Huang, X., Lin, G., Li, W., Tu, Z., Hu, H., Hu, J., & Wang, Y. (2023) · Advances in Neural Information Processing Systems 36 · doi:10.52202/075280-3398 · Cited by 128
FakeAVCeleb: A Novel Audio-Video Multimodal Deepfake Dataset
Khalid, H., Tariq, S., Kim, M., & Woo, S. S. (2021) · arXiv · doi:10.48550/arxiv.2108.05080 · Cited by 116
Audio-Visual Person-of-Interest DeepFake Detection
Cozzolino, D., Pianese, A., Nießner, M., & Verdoliva, L. (2023) · 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw59228.2023.00101 · Cited by 104
Self-Supervised Video Forensics by Audio-Visual Anomaly Detection
Feng, C., Chen, Z., & Owens, A. (2023) · 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52729.2023.01011 · Cited by 101
DeepfakeBench: A Comprehensive Benchmark of Deepfake Detection
Yan, Z., Zhang, Y., Yuan, X., Lyu, S., & Wu, B. (2023) · Advances in Neural Information Processing Systems 36 · doi:10.52202/075280-0201 · Cited by 98
Exploring Frequency Adversarial Attacks for Face Forgery Detection
Jia, S., Ma, C., Yao, T., Yin, B., Ding, S., & Yang, X. (2022) · 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52688.2022.00407 · Cited by 97 · cited 2×
“Frequency-based detectors can improve resilience to certain compression regimes but are also subject to adaptive attacks that explicitly target frequency cues [10].”
DeepfakeBench: A Comprehensive Benchmark of Deepfake Detection
Yan, Z., Zhang, Y., Yuan, X., Lyu, S., & Wu, B. (2023) · arXiv · doi:10.48550/arxiv.2307.01426 · Cited by 97 · cited 7×
“When only the manipulation algorithm changes within the same dataset, AUC degradation can be substantial for many detectors [8].”
InfluentialAVFF: Audio-Visual Feature Fusion for Video Deepfake Detection
Oorloff, T., Koppisetti, S., Bonettini, N., Solanki, D., Colman, B., Yacoob, Y., Shahriyari, A., & Bharaj, G. (2024) · 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52733.2024.02559 · Cited by 85
Tree-Rings Watermarks: Invisible Fingerprints for Diffusion Images
Wen, Y., Kirchenbauer, J., Geiping, J., & Goldstein, T. (2023) · Advances in Neural Information Processing Systems 36 · doi:10.52202/075280-2529 · Cited by 70
SynthID-Image: Image watermarking at internet scalePreprint
Gowal, S., Bunel, R., Stimberg, F., Stutz, D., Ortiz-Jiménez, G., Kouridi, C., Vecerík, M., Hayes, J., Rebuffi, S.-A., Bernard, P., et al. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2510.09263 · Cited by 70
AV-Deepfake1M: A Large-Scale LLM-Driven Audio-Visual Deepfake Dataset
Cai, Z., Ghosh, S., Adatia, A. P., Hayat, M., Dhall, A., Gedeon, T., & Stefanov, K. (2024) · Proceedings of the 32nd ACM International Conference on Multimedia · doi:10.1145/3664647.3680795 · Cited by 59 · cited 1×
“AV-Deepfake1M and its follow-on challenge infrastructure target large-scale audio-visual detection and localization [42], DeepSpeak emphasizes consented and identity-matched webcam deepfakes generated with modern voice and video engines [43], and Deepfake-Eval-2024 highlights how sharply detector…”
AV-Deepfake1M: A Large-Scale LLM-Driven Audio-Visual Deepfake Dataset
Cai, Z., Ghosh, S., Adatia, A. P., Hayat, M., Dhall, A., Gedeon, T., & Stefanov, K. (2023) · arXiv · doi:10.48550/arxiv.2311.15308 · Cited by 59
T2VPhysBench: A First-Principles Benchmark for Physical Consistency in Text-to-Video GenerationPreprint
Guo, X., Huo, J., Shi, Z., Song, Z., Zhang, J., & Zhao, J. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2505.00337 · Cited by 55
Protecting world leaders against deep fakes using facial, gestural, and vocal mannerisms
Bohacek, M., & Farid, H. (2022) · Proceedings of the National Academy of Sciences · doi:10.1073/pnas.2216035119 · Cited by 45 · cited 1×
“A directly gait-centric multimodal deepfake-detection literature remains limited [20–22].”
RingID: Rethinking Tree-Ring Watermarking for Enhanced Multi-key Identification
Ci, H., Yang, P., Song, Y., & Shou, M. Z. (2025) · Lecture Notes in Computer Science · doi:10.1007/978-3-031-73390-1_20 · Cited by 36
Audio–visual deepfake detection using articulatory representation learning
Wang, Y., & Huang, H. (2024) · Computer Vision and Image Understanding · doi:10.1016/j.cviu.2024.104133 · Cited by 28 · cited 2×
“A directly gait-centric multimodal deepfake-detection literature remains limited [20–22].”
Watch Those Words: Video Falsification Detection Using Word-Conditioned Facial Motion
Agarwal, S., Hu, L., Ng, E., Darrell, T., Li, H., & Rohrbach, A. (2023) · 2023 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) · doi:10.1109/wacv56688.2023.00469 · Cited by 28 · cited 1×
“Beyond temporal synchronization, recent multimodal deepfake detection research has begun to exploit richer behavioral and biometric correspondences between face, voice, and motion [16–20].”
Singh, S., & Dhumane, A. (2025) · MethodsX · doi:10.1016/j.mex.2025.103632 · Cited by 24
Fine-Grained Multimodal DeepFake Classification via Heterogeneous Graphs
Yin, Q., Lu, W., Cao, X., Luo, X., Zhou, Y., & Huang, J. (2024) · International Journal of Computer Vision · doi:10.1007/s11263-024-02128-1 · Cited by 23
National Institute of Standards and Technology (US) (2024) · doi:10.6028/nist.ai.100-4 · Cited by 9
Barrington, S., Bohacek, M., & Farid, H. (2024) · arXiv · doi:10.48550/arxiv.2408.05366 · Cited by 9
User guide for NIST Media Forensic Challenge (MFC) datasets
Guan, H., Delgado, A., Lee, Y., Yates, A. N., Zhou, D., Kheyrkhah, T., & Fiscus, J. (2021) · doi:10.6028/nist.ir.8377 · Cited by 8 · cited 3×
“In parallel, National Institute of Standards and Technology has supported public evaluation efforts (e.g., Open-MFC) designed to reduce bias from self-evaluation and to structure recurring, task-based benchmarking for manipulation and deepfake detection [5].”
InfluentialHow Far is Video Generation from World Model: A Physical Law Perspective
Kang, B., Yue, Y., Lu, R., Lin, Z., Zhao, Y., Wang, K., Huang, G., & Feng, J. (2024) · arXiv · doi:10.48550/arxiv.2411.02385 · Cited by 7
Artificial intelligence for deepfake detection: systematic review and impact analysis
Sunkari, V., & Sri Nagesh, A. (2024) · IAES International Journal of Artificial Intelligence (IJ-AI) · doi:10.11591/ijai.v13.i4.pp3786-3792 · Cited by 6 · cited 1×
“Representative deepfake-detection models with training, validation, and testing accuracies range from 94.2/91.8/90.5% for an RNN-based model to 99.1/97.5/97.2% for an ensemble model [48].”
InfluentialDetecting Deepfake Talking Heads from Facial Biometric Anomalies
Norman, J., & Farid, H. (2026) · IEEE/CVF Winter Conference on Applications of Computer Vision Workshops (WACVW) · doi:10.1109/wacvw68408.2026.00029 · Cited by 6
Deepfake-Eval-2024: A Multi-Modal In-the-Wild Benchmark of Deepfakes Circulated in 2024
Chandra, N. A., Lee, H., Murtfeldt, R., Qiu, L., Karmakar, A., Tanumihardja, E., Farhat, K., Caffee, B., Lee, C., Choi, J., et al. (2025) · arXiv · doi:10.48550/arxiv.2503.02857 · Cited by 5
Research works printed but not linked: 1
- European Union (2026, March 31). Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024 Laying Down Harmonised Rules on Artificial Intelligence and Amending REGULATIONS (EC) No 300/2008, (EU) No 167/2013, (EU) No 168/2013, (EU) 2018/858, (EU) 2018/1139 and (EU) 2019/2144 and Directives 2014/90/EU, (EU) 2016/797 and (EU) 2020/1828 (Artificial Intelligence Act). Official Journal of the European Union, L, 2024/1689. EUR-Lex. Available online: https://eur-lex.europa.eu/eli/reg/2024/1689/oj/eng.
Cited by
Vysotska, V., Chyrun, L., Лаврут, О., Lavrut, T., & Romavchyk, R. (2026) · Системи управління навігації та зв’язку Збірник наукових праць · doi:10.26906/sunz.2026.2.052
Deepfakes and Synthetic Media: Generation, Detection, and Governance
Gazis, A., Karypidis, E., Santamouri, K., Vavouras, T., Mastorakis, N. E., & Pappas, S. S. (2026) · Encyclopedia · doi:10.3390/encyclopedia6080165 · Cited by 5
“As such, for this reason, current best practice is defense in depth, combining [18]: 1.”
Metadata
- Authors
- Erokhin, D., & Komendantova, N.
- Affiliations
- International Institute for Applied Systems Analysis (AT)
- Corresponding author
- Erokhin, Dmitry
- Venue
- Information
- Publisher
- MDPI AG
- Published
- 2026-04-03
- Volume
- 17
- Issue
- 4
- Pages
- 347
- ISSN
- 2078-2489 (ISSN-L)
- OpenAlex
- W7149940992
- Semantic Scholar
- 5a6833f57316df3763516f7ee9ccbca2d31dfe7d
- Cited by
- 3
- References
- 26
- Citations by year
- 2026: 3
- FWCI
- 4.69
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- Information (publishedVersion, cc-by)
- IIASA PURE (International Institute of Applied Systems Analysis) (submittedVersion, other-oa)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Digital Media Forensic Detection
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Steganography and Watermarking Techniques
- Social Sciences › Sociology and Political Science › Misinformation and Its Impacts
- Keywords
- deepfakes, synthetic media, deepfake detection, content provenance, digital watermarking, media authentication, adversarial robustness, information integrity, digital trust, platform governance, legal and ethical issues, multimodal media
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.