Survival of the fairest: ethical AI in film and music
December 10, 2025
Over the past year, the entertainment world has undergone a quiet yet deep shift: the focus is moving from what AI can create to how it’s created. The discussion isn’t just about generative power anymore. Even though the UK’s legal position on using copyright-protected materials to train AI tools remains unsettled, audiences and platforms are growing more mindful of how content is produced, sparking broader conversations about AI training, consent, and compensation.
How is AI changing entertainment?
AI isn’t new to entertainment, but its role is rising as its creative potential expands. High-profile projects like Emilia Pérez and The Brutalist illustrate how AI tools can streamline production processes across film, TV, and music. Yet much of AI’s creative growth has felt like a digital land grab, prompting concerns across the industry. Companies that scrape publicly available content—images, lyrics, voices, and video clips—without explicit licensing risk infringing the rights of creators and copyright holders. We’re already seeing multiple legal battles over copyright infringement in this new AI era.
Which direction is the film industry taking?
Even with pervasive global legal uncertainty, some players are choosing a different path. AI studio Asteria exemplifies this shift with its ethically trained video models. Partnering with AI-creative innovators Moonvalley, which is known for advanced text-to-video and image-generation tools, Asteria released a model named Marey. Marey is trained only on data provided with consent and compensation to content owners, using a mix of licensed content, acquired datasets, and newly generated material to fill any gaps.
This strategy aims to give users of AI tools and the resulting content confidence that they won’t face the infringement and reputational risks associated with models trained on unlicensed data, while also supporting creators.
Asteria is already applying its technology in real production: All Heart, an animated short by Oscar-winning duo Michael Govier and Will McCormack. For All Heart, Asteria trained a custom AI model to generate production assets in the illustrator’s distinctive style and integrated these elements into the traditional animation workflow. The filmmakers praised the approach, saying it unlocked more freedom to art-direct the animation and allowed them to redirect AI-generated time savings toward realizing their artistic visions.
This example shows AI can be a complementary tool for creative direction within a transparent, rights-respecting framework that minimizes third-party infringement.
How is AI being fine-tuned in music?
The ethical AI paradigm isn’t limited to film. In music, Universal Music Group has teamed up with Stability AI, a UK-based open-source generative AI company that builds models across image, video, audio, and language domains. The collaboration has produced a suite of artist-focused AI tools, grounded in responsibly licensed datasets and direct engagement with the creative community. Central to this effort is placing artists at the heart of development and ensuring fair protection and compensation.
The music industry has long built strong IP frameworks around ownership and royalties. This shift signals a willingness to collaborate with AI developers who use transparent, lawful training methods, mirroring the direction seen in film with initiatives like Asteria.
Key trends and takeaways
Ethical AI is motivated by a dual drive: protecting creators and fostering collaboration among industry stakeholders. It’s also fundamentally commercial. Artists fiercely guard their work, and studios invest heavily in their catalogs. New AI offerings that earn trust by avoiding copyright infringement have a clear competitive edge. Transparency and governance are becoming valuable differentiators in entertainment AI. Ethical training may be essential for boosting both trust and creative freedom.
This makes sense: studios and labels face real exposure if they deploy content trained on unlicensed data, including copyright claims and reputational damage. The thread connecting Asteria’s “clean AI” initiative and Universal’s ethical partnership is that, as generative systems become production tools rather than experiments, the training methods will shape their legitimacy and perhaps their market viability. In the end, these choices may determine whether such tools survive in a rights-sensitive industry.
For entertainment companies—whether in production or distribution—AI usage within the broader supply chain could become a core strategic issue. This isn’t merely about legal risk anymore; it’s about protecting reputation, strengthening long-term positioning, and maintaining constructive relationships with creatives and rights holders.
What’s happening legally with creative AI in the UK?
The UK creative community has been vocal about the need for stronger AI oversight in the arts. In September, a broad coalition of creative groups and prominent figures, including Paul McCartney and Elton John, signed an open letter urging the government to regulate AI and copyright. The emphasis is on safeguarding creators’ upload rights and copyright protections. This growing pressure underscores the urgency of governance in the AI era and reinforces the value of ethical, licensed training models in the UK, particularly as we await the government’s updates on reforms shaping the regulatory landscape.
The debate gained further momentum after the landmark Getty Images v. Stability AI decision, which we examined in a separate piece about its implications for AI developers and rights holders. Read more here: https://www.dentons.com/en/insights/articles/2025/november/5/ai-gains-stability-uk-court-finds-no-secondary
Will ethical training models establish a lasting industry standard? It’s too soon to tell. For now, these models show that licensing, consent, and fair compensation choices could determine how innovation and authorship coexist for decades to come.
