Multiformat Unveils AI-Created Vintage Adult Films

Variety reports that a collection of AI-generated short films based on 1976 erotic magazine photo spreads debuted on the sidelines of the Cannes Film Festival, developed by Thomas Meier of Norwegian company Multiformat. Per Variety, Meier used generative AI tools to convert static spreads into full-motion video with color, synchronized sound, dialogue and voice-over. The collection streamed on Cultpix and, according to Variety, will receive a physical release on BluRay and a limited VHS edition from Klubb Super 8. Cultpix CEO Rickard Gramfors is quoted saying, "What was once considered shocking 'adult' material now seems remarkably innocent by today's standards," and frames the project as a conversation between past aesthetics and new technology, per Variety. Variety places the release amid growing curatorial interest in vintage erotica and festival programming that revisits censorship debates.
What happened
Per Variety, a collection of AI-generated short films derived from 1976 erotic magazine photo spreads premiered on the sidelines of the Cannes Film Festival, the article reports. Variety attributes the project to Thomas Meier of Norwegian company Multiformat, and says Meier used generative AI tools to transform static magazine spreads into full-motion video with color, synchronized sound, dialogue and voice-over. Variety reports the collection streamed on Cultpix and that a physical release on BluRay and a limited VHS edition will follow from Klubb Super 8. Cultpix CEO and co-founder Rickard Gramfors is quoted by Variety: "What was once considered shocking 'adult' material now seems remarkably innocent by today's standards."
Technical details
Per Variety, the technical approach used generative AI to animate archival still photography and add audio elements; the article does not specify which models, tools, or vendors were used. Variety's reporting focuses on the aesthetic and curatorial framing rather than publishing a technical workflow or model names.
Industry context
Context and significance
Editorial analysis
Companies and creators using generative models to convert archival stills into motion are part of a broader pattern where AI is applied to restoration, reanimation and creative reinterpretation of historical media. For practitioners, that pattern raises reuse, provenance and rights questions that often accompany archival dataset creation and media synthesis workflows.
Public reporting frames this Cannes-adjacent release as both an artistic statement and a provocation about how standards for nudity and sexuality have evolved. Variety situates the project alongside festival programming and institutional retrospectives that revisit censorship debates, indicating growing curatorial interest in vintage erotica and the cultural implications of automated media production.
What to watch
For practitioners
observers should look for follow-up disclosures about the models, training data sources, and rights clearances. Coverage that names specific tools or explains licensing for the underlying magazine images will materially affect legal and reproducibility assessments in archival synthesis projects.
Key Points
- 1A Cannes-adjacent debut shows generative AI being used to animate archival erotica, drawing curatorial and festival attention.
- 2Variety's reporting highlights the artistic and cultural framing rather than technical specifics, leaving model provenance and licensing opaque.
- 3For practitioners, increasing use of AI on archival media elevates provenance, rights clearance, and reproducibility as operational priorities.
Scoring Rationale
The story is notable to practitioners because it shows a cultural-institution adoption of generative AI for creative media, but it contains few technical details or industry-shaping claims. That yields solid relevance without frontier-level impact.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems

