Talking Face Generation in Socio-Technical Systems: A Systematic Review of Deep Learning Architectures, Deployment Contexts, and Governance Frameworks

Summary

Salahudeen et al. (2026) systematically reviewed 92 primary studies of talking face generation published between 2016 and 2025, covering architectures, training data, evaluation metrics and safeguards. Diffusion Transformers and Gaussian Splatting account for about 36% of studies dated 2024 or later, training data are demographically skewed and under-documented, the dominant metrics do not measure what deployment requires, and only one of the 92 systems embeds any technical safeguard while none implements consent or provenance. The authors propose the RTFG-STS Framework and three deployment-fitness metrics as research constructs requiring empirical validation.

Practical implication. Venues, funders and regulators need to make provenance, consent and demographic auditing conditions of publication and deployment for talking face generation, the authors argue.

The Problem

Talking face generation can now synthesise photorealistic, real-time speaking video from a single image and an audio clip and is entering healthcare, education, identity management and public communication, yet no prior review was at once systematic, current to 2024–2025, and comprehensive across its technical, ethical and socio-technical dimensions.

The Methodology

Systematic literature review (Kitchenham and Charters; reported to PRISMA 2020) of 92 primary studies published 2016–2025, screened from 18,343 records in six repositories over four search iterations, with a five-criterion quality assessment and qualitative narrative synthesis across six research questions.

The Core Finding

Across 92 primary studies (2016–2025), Diffusion Transformers and Gaussian Splatting account for about 36% of studies dated 2024 or later while GAN use has fallen below 3%; training data are demographically skewed and under-documented, and the dominant metrics (PSNR, SSIM, SyncNet) do not measure what deployment requires. Only one of the 92 systems embeds any technical safeguard — a non-compliant post hoc watermark — and none implements consent or provenance, even though EU AI Act Article 50 transparency obligations have applied since 2 August 2026.

The Citation

Salahudeen, R., Fonkam, M., Vajjhala, N. R., Junaidu, S. B., & Garba, A. (2026). Talking Face Generation in Socio-Technical Systems: A Systematic Review of Deep Learning Architectures, Deployment Contexts, and Governance Frameworks. Systems, 14(10), Article 1198. https://doi.org/10.3390/systems14101198 Copy APA

The summary, key findings, methods, implications and relevance notes on this page are this website’s own description of the paper, written from the published abstract and text. The official abstract and citation details are given further down.

The study in detail

What question does this paper answer?

How have deep learning architectures for talking face generation evolved since 2016, how well do their training data, evaluation metrics and safeguards match what real deployments require, and what governance would make their use in healthcare, education, identity management and public communication responsible?

What did the study find, in detail?

Synthesising 92 primary studies published between 2016 and 2025, the review finds that talking face generation has passed through five architectural generations, with Diffusion Transformers and Gaussian Splatting now accounting for about 36% of studies dated 2024 or later and GAN use below 3%. Training data are demographically skewed and under-documented, the dominant metrics (PSNR, SSIM, SyncNet) do not measure what deployment requires, and only one of the 92 systems embeds any technical safeguard — a post hoc watermark that is not C2PA-compliant — while none implements consent or provenance.

Why does it matter?

Talking face generation now produces photorealistic, real-time speaking video from a single image and an audio clip and is entering healthcare, education, identity management and public communication, yet the reviewed state of the art satisfies none of the EU AI Act Article 50 transparency obligations that have applied since 2 August 2026. The authors argue that the gap is structural rather than a failure of individual researchers — publication incentives reward benchmark optimisation alone — so venues, funders and regulators need to make provenance, consent and demographic auditing conditions of publication and deployment. The RTFG-STS Framework and three proposed deployment-fitness metrics give researchers and deployers a starting point that still requires empirical validation.

Methods and data

Standards, datasets and methods named in this study

• Guidelines for performing systematic literature reviews in software engineering Review guideline
  The review follows Kitchenham and Charters’ guidelines for systematic literature reviews.
  Kitchenham, B., & Charters, S. (2007). Guidelines for performing systematic literature reviews in software engineering (EBSE Technical Report EBSE-2007-01). Keele University and Durham University.
• PRISMA 2020 statement Reporting guideline
  The review is reported to PRISMA 2020.
  Page, M. J., McKenzie, J. E., Bossuyt, P. M., Boutron, I., Hoffmann, T. C., Mulrow, C. D., et al. (2021). The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. BMJ, 372, n71. https://doi.org/10.1136/bmj.n71
• EU Artificial Intelligence Act — Regulation (EU) 2024/1689 Regulation
  The review measures the reviewed systems against the transparency obligations of Article 50.
  Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024 laying down harmonised rules on artificial intelligence (Artificial Intelligence Act). Official Journal of the European Union.
• C2PA Technical Specification (Content Credentials) Technical standard
  The provenance standard against which the one safeguard found was judged non-compliant. The review cites C2PA v2.0; this link opens version 2.4 of the specification.
  Coalition for Content Provenance and Authenticity (C2PA). Content Credentials: C2PA Technical Specification.
• TAKE IT DOWN Act (United States, Public Law 119-12) Law
  Named among the governance instruments adopted within the review period.
  TAKE IT DOWN Act, S. 146, 119th Congress, Public Law 119-12 (2025).
• AI4People ethical framework Ethics framework
  The proposed RTFG-STS Framework adds a sixth principle to the five AI4People principles.
  Floridi, L., Cowls, J., Beltrametti, M., Chatila, R., Chazerand, P., Dignum, V., et al. (2018). AI4People—An ethical framework for a good AI society: Opportunities, risks, principles, and recommendations. Minds and Machines, 28, 689–707. https://doi.org/10.1007/s11023-018-9482-5
• VoxCeleb2 Dataset
  The training corpus that underpins 60 of the 92 reviewed studies (65.2%).
  Chung, J. S., Nagrani, A., & Zisserman, A. (2018). VoxCeleb2: Deep speaker recognition. Proceedings of Interspeech 2018, 1086–1090.
• CREMA-D Dataset
  The only corpus among those reviewed that reports ethnicity statistics; it is used in a single study.
  Cao, H., Cooper, D. G., Keutmann, M. K., Gur, R. C., Nenkova, A., & Verma, R. (2014). CREMA-D: Crowd-sourced emotional multimodal actors dataset. IEEE Transactions on Affective Computing, 5(4), 377–390. https://doi.org/10.1109/TAFFC.2014.2336244

Definitions and key terms

Talking face generation (TFG)

The synthesis of photorealistic, temporally coherent video in which a face speaks in synchrony with a given audio signal, typically from a single portrait image and an audio clip. — as used in Salahudeen et al. (2026)

Diffusion Transformer (DiT)

A diffusion model in which the usual U-Net denoising network is replaced by a transformer operating over latent tokens, so that output quality scales with model size and compute. — as used in Salahudeen et al. (2026)

Gaussian Splatting

A 3D rendering approach that represents a head or scene as learned 3D Gaussians and draws them by fast differentiable rasterisation instead of volume rendering, which makes real-time talking-head synthesis possible. — as used in Salahudeen et al. (2026)

Content provenance (C2PA)

A cryptographically signed manifest, specified by the Coalition for Content Provenance and Authenticity (C2PA v2.0), recording how a media asset was produced and bound to that asset, so that the authenticity and integrity of the claim can be verified. — as used in Salahudeen et al. (2026)

Deepfake Detectability Rate (DDR)

A metric proposed in this review: the proportion of generated frames correctly classified as synthetic by an ensemble of state-of-the-art deepfake detectors, intended to operationalise the machine-detectability required by EU AI Act Article 50(2). — as used in Salahudeen et al. (2026)

Systemic embeddedness

The sixth principle the RTFG-STS Framework adds to the five AI4People principles: the design and governance of a generative system should reference the concrete socio-technical context in which it is deployed, not the model in isolation. — as used in Salahudeen et al. (2026)

All key terms across the publications →

Abstract

Talking face generation (TFG)—the synthesis of photorealistic speaking video from a portrait and an audio signal—has passed through five architectural generations since 2016, from Long Short-Term Memory (LSTM) lip-sync models through Generative Adversarial Networks (GANs), Neural Radiance Fields (NeRFs) and Diffusion Models to Diffusion Transformers (DiT) and Gaussian Splatting, and now enters healthcare, education, identity management and public communication in real time. This systematic review synthesises 92 primary studies (January 2016–December 2025), selected from 18,343 records across six repositories under Kitchenham’s guidelines and PRISMA 2020, addressing architecture evolution, dataset bias, evaluation metrics, paradigm shifts, ethical safeguards and socio-technical deployment. DiT and Gaussian Splatting account for roughly 36% of studies dated 2024 or later, and GAN use has fallen below 3%; training corpora are demographically skewed and under-documented, and the dominant metrics (PSNR, SSIM, SyncNet) do not measure what deployment requires. Only one of the 92 systems embeds any technical safeguard, a non-compliant post hoc watermark, and none implements consent or provenance, although the EU AI Act, the TAKE IT DOWN Act, China’s labelling Measures and C2PA v2.0 were all adopted within the review period. We propose three deployment-fitness metrics and the RTFG-STS Framework as research constructs requiring empirical validation.

Abstract as published in Systems under a Creative Commons licence.

Keywords: talking face generation; socio-technical systems; deepfake governance; systematic literature review; responsible AI; human-AI interaction

Limitations

• Restricting the review to English-language publications underrepresents Chinese-language venues, although Chinese institutions are among the field’s most active contributors.
• Studies report incompatible metric subsets on incompatible test sets, so no meta-analysis was possible and differences in quality between architectures cannot be quantified from the published literature.
• The findings on ethical safeguards and socio-technical deployment (RQ5, RQ6) rest on a thinner evidence base that also draws on regulatory instruments, industry reporting and adjacent literature gathered purposively rather than under the review protocol.
• The three deployment-fitness metrics and the RTFG-STS Framework are research constructs that have not been validated against human raters or a detection benchmark, and are not compliance criteria.
• The regulatory analysis reflects the position as at September 2026 and should be re-checked against later EU AI Office guidance.

Practical implications

• The authors argue that, because publication incentives reward benchmark optimisation alone, venues, funders and regulators need to make provenance, consent and demographic auditing conditions of publication and deployment.
• For deployers, the reviewed state of the art satisfies none of the EU AI Act Article 50 transparency obligations that have applied since 2 August 2026.
• Researchers and deployers can use the RTFG-STS Framework and the three proposed deployment-fitness metrics as a starting point, although these still require empirical validation.
• For researchers and deployers, the dominant metrics do not measure what deployment requires: PSNR, SSIM and SyncNet measure none of identity preservation, cross-cultural naturalness, trustworthiness or detectability.

Drawn from Salahudeen et al. (2026), Systems, 14(10), Article 1198.

When this research may be relevant

This review may be relevant to researchers surveying talking face or talking head generation, audio-driven portrait animation and deepfake synthesis, and to work on dataset bias, demographic representation and evaluation metrics for generative video models. It also speaks to AI governance and regulation of synthetic media — content provenance (C2PA), watermarking, consent and EU AI Act Article 50 transparency obligations — and offers a socio-technical systems view of deploying generative avatars in healthcare, education, identity verification and public communication.

Research topics addressed: talking face generation; talking head generation; deepfakes; generative AI; deep learning; computer vision; diffusion models; Diffusion Transformers; Gaussian splatting; neural radiance fields; dataset bias; evaluation metrics; AI governance; EU AI Act; content provenance (C2PA); identity management; socio-technical systems; systematic review

How to cite

APA 7

Salahudeen, R., Fonkam, M., Vajjhala, N. R., Junaidu, S. B., & Garba, A. (2026). Talking Face Generation in Socio-Technical Systems: A Systematic Review of Deep Learning Architectures, Deployment Contexts, and Governance Frameworks. Systems, 14(10), Article 1198. https://doi.org/10.3390/systems14101198Copy APA citation

MLA 9

Salahudeen, Ridwan, et al. “Talking Face Generation in Socio-Technical Systems: A Systematic Review of Deep Learning Architectures, Deployment Contexts, and Governance Frameworks.” Systems, vol. 14, no. 10, 2026, article 1198. https://doi.org/10.3390/systems14101198.Copy MLA citation

IEEE

R. Salahudeen, M. Fonkam, N. R. Vajjhala, S. B. Junaidu, and A. Garba, “Talking Face Generation in Socio-Technical Systems: A Systematic Review of Deep Learning Architectures, Deployment Contexts, and Governance Frameworks,” Systems, vol. 14, no. 10, Art. no. 1198, Sept. 2026, doi: 10.3390/systems14101198.Copy IEEE citation

@article{salahudeen2026talking,
  title = {Talking Face Generation in Socio-Technical Systems: A Systematic Review of Deep Learning Architectures, Deployment Contexts, and Governance Frameworks},
  author = {Salahudeen, Ridwan and Fonkam, Mathias and Vajjhala, Narasimha Rao and Junaidu, Sahalu Balarabe and Garba, Aliyu},
  journal = {Systems},
  volume = {14},
  number = {10},
  pages = {1198},
  year = {2026},
  publisher = {MDPI},
  doi = {10.3390/systems14101198},
  url = {https://doi.org/10.3390/systems14101198}
}

TY  - JOUR
TI  - Talking Face Generation in Socio-Technical Systems: A Systematic Review of Deep Learning Architectures, Deployment Contexts, and Governance Frameworks
AU  - Salahudeen, Ridwan
AU  - Fonkam, Mathias
AU  - Vajjhala, Narasimha Rao
AU  - Junaidu, Sahalu Balarabe
AU  - Garba, Aliyu
JO  - Systems
PY  - 2026
DA  - 2026/09/24
VL  - 14
IS  - 10
SP  - 1198
PB  - MDPI
SN  - 2079-8954
DO  - 10.3390/systems14101198
UR  - https://www.narasimharao.net/research/talking-face-generation-socio-technical-systems-systematic-review/
LA  - en
ER  - 

Links and files

• DOI: https://doi.org/10.3390/systems14101198
• Publisher’s page: https://www.mdpi.com/2079-8954/14/10/1198
• Full text (PDF): https://www.mdpi.com/2079-8954/14/10/1198/pdf
• This page as text: https://www.narasimharao.net/research/talking-face-generation-socio-technical-systems-systematic-review/paper.md
• Permanent address of this page: https://www.narasimharao.net/research/talking-face-generation-socio-technical-systems-systematic-review/