Game-changers: Hollywood’s deal with WGA rewrites AI rules

The tentative deal announced last week by the Writers Guild of America includes many industry-specific aspects, such as the size of writers’ rooms and improved residuals for streaming. But everyone from autoworkers to white-collar middle managers should be paying very close attention to how this deal was achieved — because it sets a monumental precedent for labour relations in a digital future.

Unlike bread-and-butter issues like wages, benefits and the terms and conditions of employment — topics over which management, according to labour law, must negotiate “in good faith” — technology, like business strategy, constitutes managerial prerogative. This means that the use of technology, such as artificial intelligence, exists in a bargaining gray area. If workers bring it up, employers can either open a negotiation or reply with a simple “No, thanks.” But the WGA deal put AI squarely on the table.

In the past, management would often make nearly all technology-related decisions before negotiations even began. Workers and their unions were excluded from early conversations about technology — including ones that could open up a range of issues around use and deployment that would benefit both workers and employers. The only mandated negotiations were around the technologies’ impact on terms and conditions such as wages. This dynamic may help explain why, at the beginning of the WGA negotiations, many believed, or at least publicly stated, that securing any constraints at all on the use of AI was a pipe dream.

It wasn’t. The WGA contract establishes a precedent that an employer’s use of AI can be a central subject of bargaining. It further establishes the precedent that workers can and should have a say in when and how they use artificial intelligence at work.

It may come as a surprise to some that the WGA apparently never wanted, nor sought, an outright ban on the use of tools like ChatGPT. Instead, it aimed for a more important assurance: that if AI raises writers’ productivity or the quality of their output, guild members should snare an equitable share of the performance gains. And the WGA got it.

How did it achieve this? In this case, the parties agreed that AI is not a writer. The studios cannot use AI in place of a credited and paid guild member. Studios can rely on AI to generate a first draft, but the writers to whom they deliver it get the credit. These writers receive the same minimum pay they would have had they written the piece from scratch. Likewise, writers can elect to use AI on their own, when a studio allows it. However, no studio can require a guild member to use AI.

These negotiations were likely aided by the fact that studios have their own concerns about AI Material generated by AI cannot be copyrighted, which presents several challenges to studios, since they typically own the copyright on material from writers they’ve hired. By assuring that human writers will be involved in any writing that also involves AI, the studios start to insulate themselves from those copyright concerns.

In the past, when labour has sought to simply resist or impede technological change, it’s been completely run over. Ask any of the 6,000 hot-metal typesetters, compositors and other workers whom News International summarily dismissed in 1986 when Rupert Murdoch secretly transferred newspaper production from London’s famous Fleet Street to the company’s state-of-the-art computerized facility in the London Docklands. Rather than negotiating a more gradual transition, along with buyouts and rules around worker reskilling and reassignment, the union dug in — and lost. They walked away with nothing. The event catapulted Murdoch to international fame without costing him a single day of production or distribution on any of his papers.

The WGA may have taken a lesson from more successful tactics, such as those of the International Longshore and Warehouse Union in 1960. Their leader, Harry Bridges, convinced a reluctant rank-and-file that they would have to face the reality of port containerization, meaning many jobs would disappear as shipping containers, and the attendant technology, became more prevalent. Instead of opposing this implementation, Bridges negotiated wage and job guarantees for most workers and generous benefits for those displaced. Bridges’s strategy was itself informed by the United Mine Workers’ Mechanization Agreement in 1950. Negotiated by John L. Lewis, it encouraged the mechanization of large mines and the windfall returns it would generate — not just for the mine owners but also for the rank-and-file workers and retirees.

The WGA’s negotiating committee recognized that AI had to be on the table, and not simply as something to be resisted or refused. In the absence of any contract language on AI, employers could have gone ahead and deployed ChatGPT or other large language models in one of two ways: Studios could have encouraged guild members to use it as a productivity-enhancing tool, but without attendant compensation. They also could have circumvented WGA members altogether. Instead, this deal guarantees a contractually mandated context in which AI can be utilized — one that benefits, rather than impedes or replaces, the workers.

Much of what we understand in the intersection of collective bargaining and technological change comes from heavy industry; after all, unions made their name protecting blue-collar, industrial workers from workplace dangers and deteriorating terms and conditions of employment. So the approximately 85 percent of American workers who are not employed in blue-collar jobs can be forgiven for failing to see how unions might protect them. After all, a college education or the donning of a pressed suit once served as a prophylactic to the economic perils of technological change.

But AI is coming for workers in every sector, no matter their academic pedigree or sartorial choices. Now the WGA has delivered a gift to future union negotiators. It’s illuminated an approach to negotiating around technology, and demonstrated the ways in which a white-collar rank-and-file can leverage labour solidarity toward the shared benefit of both management and employees.

Litwin is an associate professor of industrial and labour relations at Cornell University