Trend AnalysisArts & Design

Deepfake Technology in Film and Visual Arts: Creative Tool or Existential Threat?

Deepfake technology enables filmmakers to de-age actors, resurrect historical figures, and create impossible visual effectsโ€”but the same technology powers misinformation at scale. The dual-use challenge demands both creative innovation and robust detection systems.

By Sean K.S. Shin
This blog summarizes research trends based on published paper abstracts. Specific numbers or findings may contain inaccuracies. For scholarly rigor, always consult the original papers cited in each post.

Why It Matters

Deepfake technologyโ€”the use of deep learning to generate synthetic visual and audio mediaโ€”occupies a unique position in the arts: it is simultaneously one of the most powerful creative tools and one of the most dangerous misinformation technologies ever developed. In filmmaking, deepfakes enable de-aging (Martin Scorsese's "The Irishman"), posthumous performances (Peter Cushing in "Rogue One"), and visual effects that would have been impossible or prohibitively expensive with traditional techniques. In the wider world, the same technology powers non-consensual intimate imagery, political disinformation, and financial fraud.

This dual-use nature makes deepfake technology a critical case study for how the arts engage with technologies that have both creative and destructive potential. The research landscape reflects this tension: one stream focuses on creative applications, another on detection and defense.

The Science / The Practice

Creative Applications in Film Production

Mittal et al. (2024), with 3 citations, provide the most direct examination of deepfakes as a legitimate filmmaking tool. Their paper explores how face-swapping, de-aging, and voice synthesis technologies can enhance visual effects production. The key argument is that deepfake technology is a natural evolution of the VFX pipelineโ€”digital compositing, CGI characters, and motion capture were all controversial when introduced but are now standard. The paper documents use cases where deepfakes reduce production costs by orders of magnitude compared to traditional VFX, making high-quality visual effects accessible to independent filmmakers and smaller studios.

Detection Benchmarks for Social Media

Batra et al. (2025), with 2 citations, address the other side of the equation by introducing SocialDF, a benchmark dataset and detection model specifically designed for social media platforms. Their research acknowledges that while deepfakes have legitimate creative applications, the primary distribution channel for harmful deepfakes is social media, where compression artifacts and platform-specific processing make detection more difficult. The benchmark provides standardized evaluation criteria for detection systemsโ€”a critical need in a field where claims of detection accuracy are often based on non-representative test sets.

Cultural Tourism and Creative AI

Yan et al. (2025) explore an unexpected application domain: the integration of creative AI and film visual effects technology in cultural tourism. Their work demonstrates how deepfake-adjacent technologiesโ€”virtual actors, historical scene reconstruction, and interactive AI charactersโ€”can create immersive tourism experiences that bring historical sites to life. A tourist visiting an ancient site could interact with AI-generated historical figures who explain the site's history in their "own" voices and appearances. This application highlights the creative potential of synthetic media when deployed in controlled, transparent contexts.

Blockchain-Based Authenticity Verification

Sakthivel et al. (2025) propose a combined AI and blockchain approach for deepfake detection, addressing a fundamental limitation of AI-only detection systems: even the best detection models can be fooled by adversarial attacks. By anchoring media authenticity on blockchainโ€”recording provenance, creation metadata, and modification history in a tamper-resistant ledgerโ€”the system provides a complementary defense layer. This approach is particularly relevant for the film industry, where establishing the authenticity and provenance of visual assets has legal and contractual implications.

Deepfake Technology: Creative vs. Defensive Applications

<
ApplicationDomainOpportunityRisk
De-aging / resurrectionFilm VFXImpossible scenes become possibleConsent of deceased performers
Face-swappingFilm / entertainmentCost reduction, creative flexibilityNon-consensual misuse
Historical reconstructionCultural tourismImmersive heritage experiencesHistorical inaccuracy
Voice synthesisAudio productionMultilingual dubbing, accessibilityVoice identity theft
Detection systemsPlatform safetyMisinformation defenseArms race with generation
Blockchain provenanceMedia industryTamper-resistant authenticityAdoption barriers

What To Watch

The emerging regulatory landscape will shape how deepfake technology develops in the arts. The EU AI Act classifies deepfakes as high-risk when used without disclosure, while China requires watermarking of all AI-generated content. Watch for the development of industry-standard provenance systems (such as the C2PA standard) that embed creation metadata into media files, and for the film industry's evolving labor agreements around synthetic performancesโ€”the 2023 SAG-AFTRA strike was partly driven by concerns about AI replication of actors' likenesses.

Explore related work through ORAA ResearchBrain.

References (4)

[1] Mittal, S., Joshi, M., & Vats, P. (2024). Virtual Illusions: Unleashing Deepfake Expertise for Enhanced Visual Effects in Film Production. IEEE ICRITO 2024.
[2] Batra, A., Khemani, J., & Gumber, A. (2025). SocialDF: Benchmark Dataset and Detection Model for Mitigating Harmful Deepfake Content on Social Media Platforms. ACM Proceedings.
[3] Yan, P., Li, Q., & Ma, A. (2025). Exploration of the Integrated Application of Creative AI and Film Visual Effects Technology in Cultural Tourism. ACM Proceedings.
[4] Sakthivel, K., Harine, N., & Chandra Mohan, G. (2025). Towards Tamper-Resistant Digital Media: An AI and Blockchain Approach for Deepfake Detection. IEEE ICECA 2025.

Explore this topic deeper

Search 290M+ papers, detect research gaps, and find what hasn't been studied yet.

Click to remove unwanted keywords

Search 7 keywords โ†’