By Alistair Cope and Paul Carey Jones
Not long ago the idea that AI could write a scene, score a show, build a set in 3D, or conjure a convincing human-like voice would have felt like a teaser for some far-off future. That future has arrived, and it is already reshaping the arts.
Unions are fighting for consent and fair use. Creatives in theatre, film, TV, games and voice work are pushing for rules that protect the art we make and the people who make it.
Spotify says it removed a staggering 75 million “spam tracks” in the last year, showing how quickly automated content can flood a market if guardrails are not in place. Meanwhile, streaming services also face the issue of paying for automated streaming of tracks by bots, raising the bizarre possibility of a whole industry where music is generated by and listened to entirely by computers.
In the UK, the Make It Fair campaign has become a rallying point for those whose incomes are threatened. It was launched across the creative industries to oppose proposals that would let AI companies train on artists’ work without permission or payment. Think of it as the debate that decides whether your script, score, voice or design can be used to teach a model by default. Make It Fair argues for consent, credit and pay.
As 2025 has progressed, the training of AI systems using the arts (film, music, writing) has gained more and more headlines, with high-profile figures such as Paul McCartney and Elton John criticising the proposed overhaul of the UK copyright system, and calling for new rules to prevent tech companies from riding “roughshod over the traditional copyright laws that protect artists’ livelihoods”.
None of this should overlook the grim truth that the recorded music industry is already broken, and that it was the industry itself that broke it. The number of musicians who rely on Spotify’s 0.002p per track for any significant part of their income is vanishingly small. Even before the advent of AI, the question of how the next Paul McCartney or Elton John could possibly emerge in the current set-up was an existential one. Perhaps AI-generated listeners streaming AI-generated tracks, with no humans involved at any stage of proceedings, is merely the absurdist endgame of the music industry eating itself.
Performing arts leaders have been unusually unified on the question of AI and artistic copyright. In March, directors of major organisations warned that the proposed approach would undermine a fragile freelance ecosystem that depends on copyright to make a living. Their message was clear. If policy tilts away from consent, the people most exposed are, once again, the freelancers who are the backbone of the creative industries.
Unions have also been vocal, with Equity’s ‘Stop AI Stealing the Show’ campaign setting out why opt-outs do not work in practice, and why performers need enforceable rights over their voices and likenesses. They have paired that with a practical AI Toolkit so freelancers can recognise risky clauses, negotiate better terms, and protect digital doubles at work. The TUC has since backed stronger protections and personality rights.
If music is the canary in the coalmine, screen is the klaxon. Synthetic performers and voice clones are no longer thought experiments. They are turning up at festivals and on demo reels, and they are raising the same questions we face on stage: consent, compensation, creative control.
Just last week, an AI actor named “Tilly Norwood” was unveiled at the Zurich Film Festival, prompting outraged responses across the creative industries. Tilly was promoted as a new kind of performer and was reportedly being signed by talent agents. Equity swiftly condemned the move, saying the likeness and performances of real actors may have been used without consent, calling it part of a “Wild West” of unregulated technology. The union demanded transparency, consent and fair pay, insisting that “AI-generated performances do not appear from thin air, they are made by digitally imitating real work made by real people.”
The Society of London Theatre joined the call, warning that while theatre will continue to evolve and embrace technology, its future must be safeguarded so artists are fairly rewarded and human creativity is never replaced. Their co-chief executive Claire Walker said that theatre’s magic lies in the shared, real-time connection between performers and audiences, a connection no screen or algorithm can replicate.
The timing of the Norwood controversy could not be more striking. Only this week, Zelda Williams, daughter of Robin Williams, had asked people to stop using AI to recreate her late father’s voice and likeness. She said such imitations were “deeply disturbing” and stripped away the humanity of the original performer. Her words echoed what many in the live arts are feeling: that technology is racing ahead of ethics, consent and law.
'Audiences do not come to a theatre for an infinitely repeatable file. They come for a night that only happens once.'
As the industry explores these new tools, we also need to face their environmental cost. Every AI-generated image, video or 3D render consumes large amounts of electricity and water through the data centres that power them. The carbon footprint of training and running large models is vast, with one high-end image sometimes using as much energy as boiling a kettle a dozen times - and generating text is, perhaps counterintuitively, even more carbon-intensive. For an industry that is working hard to make its practice greener and follow initiatives such as the Theatre Green Book, this matters. If we are serious about sustainability, AI must be part of that conversation. Environmental responsibility has to be a core value of our industry, not just a passing fad. We need to mean what we say. Creative innovation cannot come at the expense of the planet, and any adoption of AI in the arts needs to be matched by an awareness of its climate impact and a commitment to using it responsibly.
So where does live theatre sit in all of this? Paradoxically, both close to the fire, and uniquely placed to remind people what cannot be faked.
Theatre has already experimented with AI as a collaborator and a tool. The RSC’s Dream blended live motion capture with a responsive virtual world, which opened up a new relationship between audience and performance. The Young Vic’s AI invited GPT into the rehearsal room to generate text that actors instantly embodied on stage.
Recent coverage has shown how AI is creeping into stagecraft, from drafting designs to automating elements of sound and light, and how captioning and audio description are being transformed by better speech recognition. Currently, none of this necessarily replaces a human presence, but it does highlight areas of human work which need protecting.
For freelancers, the risks and the possibilities sit side by side. On one hand, scanning on set without informed consent has already been flagged by Equity, and opt-out proposals unfairly shift the burden onto individual workers who are least able to police how their work is scraped. On the other, smart use of tools can help a one-person company do the dull bits faster, help budget, refine a pitch, or caption a performance in multiple languages. The line to hold is simple in principle and hard in practice: consent first, clear credit, fair pay, and human oversight where it counts.
There is also a cultural truth that is easy to forget in the noise. Audiences do not come to a theatre for an infinitely repeatable file. They come for a night that only happens once. When the script breathes, when a performer misses and then finds it, when a room laughs together and then goes quiet together, that is the value. The idea that people will seek out liveness as a relief from the synthetic is not wishful thinking. It is already visible in how artists, venues and festivals talk about what makes stage work different, and in the projects that explore technology without letting it become the show.
What can we as freelancers do next? Stay informed about the policy process, especially on copyright and data mining. Back campaigns like Make It Fair. If you are a performer or creative, read Equity’s AI Toolkit and push for clear contract language about training, replicas and reuse. If you commission or produce, adopt human-centred policies now, so your freelancers do not pay the hidden cost later. And keep experimenting, but on our terms. The tools are not inherently good or bad. The rules we set and the values we hold will decide whether AI becomes a lever for access and productivity or another way to extract value from the least protected people in the room.
Clearly the arts are not immune to the dangers of AI: there is a real risk of being flooded with synthetic content that mimics the surface of what we do. Artists in TV, film, and music are all sounding the alarm, but theatre’s answer cannot be to look away. Our answer should be to show our workings in public, to insist on consent and credit, and to protect the things that make audiences feel something real. And all the while, to maintain a healthy scepticism about whatever the technological snake oil of the moment might be. The humans who create theatre professionally must be cherished and protected, and we should push back hard against any use of AI to undermine the viability of human careers.
At the end of the day, live theatre is about a shared space, a shared humanity, and the connection between human beings. An auditorium full of bots watching an AI-generated show makes no sense. The human bit is the point.
