From casting couch to final cut: how AI (Artificial Intelligence) is putting actors’ and performers’ digital rights at risk. From auditions to audiobooks, we unpick the story behind the strike.
In less than a decade, the way we create and consume media has changed beyond all recognition and with it the way an actor’s voice, image or even an entire performance can be manipulated and re-used.
Streaming services and on-line gaming are ubiquitous and, whilst they provide huge opportunities for producers and actors alike, recent technological developments present a real threat to the integrity and economic viability of the already precarious creative professions, from actors to casting directors, screen writers to composers.
For decades, the entertainment industry in the UK and USA has relied on a complex and finely balanced web of union-negotiated collective contracts, industry practice, agreed fee rates, goodwill and negotiated deals to manage the performance rights (including intellectual property rights) and associated remuneration of actors. The pay structure for residuals and royalties, created in the 1960s, means that performers benefited from a regular cheque in the post when material was shown, often decades after their original performance.
Fast forward to 2023 and generative AI is the talk of Hollywood. The structure of the streaming industry and the ability of AI to manipulate and reproduce a performer’s uniquely personal characteristics has thrown the creative world into disarray. The consequences, intended and unintended, are vast and the industry is just waking up to the issues caused by AI as they affect copyright through to contracts, and even screen writing and casting as well as the larger question of economic fairness.
TLD Client Legal Director, Sophie Freeman, is an Intellectual Property expert and legal counsel to both actors and agents. She unpacks the arguments.
One-sided contracts push actors and writers to strike.
Producers of streaming content invest huge amounts in long-running series that cost as much a full-length film but are at the mercy of their subscribers, and global pandemics! They want to reduce their risks whilst having full freedom and option to manage, edit, reproduce, extend, re-show, stop, re-start, sell, and reap the full and total economic rewards of their productions. AI has presented the producers with an almost irresistible tool in this respect, giving them the potential to re-use, re-create, adjust and copy original performances without the need to re-engage with the performers – which can be messy and expensive.
At the same time, a number of the industry-standard contracts in both the US and UK are “expiring”. In many areas the vacuum is being filled by the producer’s own very complex one-sided contracts, many of which pay talent a one-time “episodic” fee, keep actors’ careers on hold for years using multiple season “options” and provide for no or little on-going income from royalties and residuals.
At the same time, they require actors to fully assign all rights in the performance to the producer to agree that:
“the producer has the right to use, and grant others the right to use, the performance (including unincorporated footage) in various programs and other formats (including in connection with merchandise and commercial tie-in related to the Series), as determined by Producer in its sole discretion, in any and all media now known and hereafter devised, throughout the universe in perpetuity.”
When “all media” includes AI, the creative industries have a problem and this sentence (hidden in a single paragraph of over 1800 words) cannot be negotiated by any but the most powerful actors. The entertainment industry is, and always will be, a buyer’s market. The opportunity for misuse and unfair exploitation of an individual’s image (image and voice) is obvious – and threatens the privacy and data protection right of the individual. At the same time, the wider effects of the contracts means that most actors and writers cannot make a viable income from their performances and are even limited from getting any other work whilst under contract to a producer.
This has resulted in the extreme effects we see in the USA where the key unions for writers (WGA) and for television and radio artists (SAG-AFTRA) have been forced to call strikes. The UK unions are supporting these strikes but have fallen short of calling their own members to strike. There are other key demands but the fall-out from the use and misuse of AI is key.
Sophie says,
“Royalties and residual payments need to be restructured and future-proofed to enable performers to receive fair payment for the work they have done.
As legal advisors to the industry, we are campaigning for less complex, fairer contracts that recognise the integrity of the individual, properly value the digital rights of performers and pay talent appropriately for their work. With individual contracts for even small roles exceeding 60 pages and individual clauses spanning multiple pages – there is much to challenge.”
Audiobooks: are we listening carefully?
The use of AI to create ‘new’ voices has the potential to democratise audiobook creation – making it accessible for everyone – self publishers take note. But does the reduced cost of ‘synthetic narration’ come at a price for professional voice artists?
Clarity is key, says Sophie. “Providing recordings clearly state whether an audiobook is narrated by a human or synthetic voice, then publishers are being transparent with listeners and fair to voice artists. “
SAG-AFTRA is on the case, working with audiobook creators such as Amazon to negotiate fair contracts and writing a new rulebook preventing actors voices being used to ‘train’ AI systems or to create digital replicas.
Generative AI, casting, and life beyond the last scene.
Generative AI is already changing the way the entertainment industry casts and captures performances. Virtual auditions are becoming more common with AI being used to test compatibility and on-screen appeal, analysing large datasets to identify the best talent to create a box office hit.
Similarly, ‘deep fake technology’ can be used to manipulate existing footage or even fabricate performances that mimic existing actors without their involvement or consent.
Sophie’s comments,
“The concerns here are if an artist’s image is used without their consent to ‘test’ suitability for a role, what are the data protection implications? Is consent required and if not, where do we draw the line?
Test cases are now emerging in the US which seek to determine the rights of both individuals and companies to ‘opt-out’ of having their data used to train AI models. This is a challenging area as most will not know when and even if their data has already been used.”
Deep fake technology and the concept of ‘synthetic performances’ pose an even bigger threat. Clear labelling (for ‘AI generated’ images and sounds) is only part of the solution.
“Ultimately,” says Sophie “it comes back down to the contract: on the one hand protecting actors’ rights through fair terms and conditions, and on the other, in finding agreement on licensing fees that recognise the digital rights of performers. We will be busy!”
