The International Alliance of Theatrical Stage Employees (IATSE), the union representing below-the-line workers in the entertainment business, last week issued a list of “core principles” for the use of artificial intelligence technology in stage and screen production. Among them is the need to develop “clear definitions that categorize various types of so-called ‘artificial intelligence’ in order to advance enforceable legislative and collective bargaining oversight. IATSE will also urgently identify which crafts and Local Unions are most at-risk of being affected in the immediate future.”
While not currently threatening a work-stoppage, IATSE’s concern over the potential impact from AI puts its technicians and craftspeople in the same camp as their above-the-line brethren in the Writers Guild of America (WGA), who have been on strike since early May. largely over the scope and manner or AI use in the creative side of film and TV production. Their fellow above-the-line workers in the Screen Actors Guild (SAG-AFTRA) have gone into overtime in their contract negotiations with film and TV producers. But there, too, AI use is among the remaining sticking points.
For their part, the same studios the guilds are at odds with, are themselves raising alarms over the use of their catalogs of copyrighted material to train generative AI models and demanding such use be licensed and paid for. So, too, are their counterparts in the music and publishing industries.
Yet, for all their current differences and disagreements, the parties on either side of the table share an abiding fear of repeating their own past mistakes in dealing with technological transitions.
Apart from the use of AI, the other major sticking point in both the Writers Guild’s and Actors Guild’s contract talks with the studios is streaming residuals. In previous contract negotiations, both guilds agreed to accept a lower percentage of revenue from the distribution of their works on streaming platforms compared with the residual payments they receive from more traditional forms of exploitation.
They agreed to those terms in part because, at the time, streaming was not yet the dominant distribution channels for both film and television that it would become, and the studios insisted they needed financial flexibility to invest in the nascent, ancillary business. It was only after the ancillary channel became a primary one that writers and actors realized they’d taken the short-money. Worse, they had suffered a similar fate in the home video and DVD markets, which grew into cash cows for the studios.
Now twice bitten, they are three times shy over letting another new technology monster in the backdoor for failing to foresee the danger.
Similarly, music artists have seen their earnings from recordings plunge in the streaming era as all-you-can-eat licensed access has overtaken higher-royalty CD sales, even as their labels and publishers enjoy robust earnings and rising share prices. While they mostly lack the collective bargaining muscle of the Hollywood guilds, individual artists as well as groups like the Music Artists Coalition and Songwriters of North America have been alert to the new dangers of AI voice cloning and seeing their already meager royalty income diluted further by a flood of AI generated tracks.
Commercially published authors have likewise not benefitted greatly from the introduction of e-books, as lower list prices and royalty terms have meant lower payouts to writers while the marginal increase in unit sales has not been enough to make it up in volume.
Across the table meanwhile, a similar history haunts the corporate suites. When Netflix launched the over-the-top subscription streaming revolution in 2007, the movie and TV studios were only too happy to license their old, mostly moldering libraries to the new outlets. They only belatedly came to realize, a decade or so later, that other people had built billion dollar businesses on the backs of the studios’ own content even as their own share prices languished.
Since then, the studios have poured billions of their own money into launching their own streaming services in a so-far unsuccessful pursuit of Netflix-like share-price multiples. That experience has left them wary of giving anything away in the latest, AI-driven technology transition.
Two and a half decades ago, the record companies watched in horror as Napster and its ilk laid waste to the CD business. So, when Apple came along offering to sell authorized, copy-protected MP3 files through the newly launched iTunes store, it was viewed by the industry as a savior and the labels leapt at the chance to license their catalogs to Steve Jobs & Co. It was only later that they discovered, just as the studios later would, they had enabled Apple to capture the lion’s share of the music listening audience using the labels’ own catalogs and could now dictate the terms of sale.
The painful memory of losing control of their own margins is now fueling the labels’ current determination to entrench a system for licensing the use of their content by generative AI models on their own terms.
The major publishing houses have largely avoided the technology-transition calamities of the studios and record labels, due mostly to the book business so far being largely spared much of the technology-driven upheaval of other media sectors. But the publishers, along with the Authors Guild, suffered major technology setbacks in the HaithiTrust and Google Books cases in which courts found the unlicensed copying of tens of thousands of copyrighted books to create a searchable digital database to be a fair use.
Like the record labels, book publishers are now determined to avoid a repeat of that gloomy history with the use of their works to train generative AI models.
In short, nearly all the parties involved in the various negotiations and discussions around generative AI choked on their first bites at the digital apple, and are determined chew carefully and deliberately this time around. That’s not a good recipe for a quick resolution to any of those meals.