In SAG-AFTRA Ratification Vote, AI Is the Biggest Source of Contention

After a 118-day strike, 160,000 SAG-AFTRA members are voting on whether to ratify a new agreement. AI has emerged as the key source of division, with some members unsatisfied that a ban wasn’t on the table.

SAG-AFTRA members and supporters chant outside Paramount Studios on day 118 of their strike against the Hollywood studios on November 8, 2023 in Los Angeles, California. (Mario Tama / Getty Images)

It didn’t take long after the Screen Actors Guild–American Federation of Television and Radio Artists (SAG-AFTRA) announced that it reached a tentative agreement with the Alliance of Motion Picture and Television Producers (AMPTP) for signs to emerge that members would need to read the fine print. While the union’s negotiating committee unanimously voted to approve the deal on November 8 following a 118-day strike, it next went to the national board, where the vote was 86 percent in favor, rather than unanimous.

Such dissension, particularly after a lengthy strike, suggested that the three-year agreement’s ratification process wouldn’t go as smoothly as that of the Writers’ Guild of America (WGA), which formalized its tentative agreement (TA) on October 9, with 99 percent of votes favoring ratification.

The initial message sent to the membership following the November 8 vote touted the TV/Theatricals TA’s $1 billion in new wages and benefit plan funding, as well as the “unprecedented provisions for consent and compensation that will protect members from the threat of AI [artificial intellegence],” a first-ever “streaming participation bonus,” and substantial raises in the union’s pension and health caps. The next day, the union called the agreement “revolutionary.”

On wages, the TA includes a 7 percent salary increase upon ratification, followed by 4 percent in July 2024 and 3.5 percent the following year. Background performers will see an 11 percent hike effective November 12. While that is not the 12 percent the union initially proposed, it is significantly higher than the studios’ 5 percent counter, representing a break in the pattern established by the AMPTP.

As for the streaming bonus, it amounts to $40 million in additional compensation. The union explained that the “majority of that compensation will be paid to actors on programs meeting certain viewership requirements. The remaining money will be distributed to other actors working on those streaming platforms through a new, jointly-trusteed distribution fund.” Those bonuses will be doled out based on data provided by the streaming services on the total number of hours streamed for high-budget streaming programs. That represents a major compromise from the union’s initial proposal: 2 percent of streaming-service revenue, later dropped to 1 percent, then changed to a per-subscriber fee. Streamers called those proposals a “bridge too far,” refusing to budge.

“I felt like, is this a win or a loss?” SAG-AFTRA president Fran Drescher told the New York Times of the compromise. “But we’re getting the money. We opened a new revenue stream. What matters is that we got into another pocket.”

The Dispute Over AI

In the weeks since the union announced the TA, members have focused squarely on one area of the contract: AI. Some performers insist the TA has unacceptable loopholes and that they’re voting “no” on ratification. Balloting continues through December 5.

Shaan Sharma and Anne-Marie Johnson, two of the nine dissenters on the eighty-member national board, explained to Variety that the problem of generative AI was central to their “nay” votes, though not the only cause of their reservations.

“There should be no AI. Only human beings should be used in what we create for public consumption,” Johnson said. “Without staving off AI, everything we achieved is for naught. It’s a waste of time.”

“Consent, in the context of this agreement, is tyranny,” said Matthew Modine, another member of the national board who opposes the deal. The Stranger Things actor’s statement may carry particular weight, as it was Modine who ran against Drescher in the union’s hard-fought 2021 leadership elections. In a lengthy statement, Modine argued for rejecting the agreement and returning to the bargaining table to strengthen the AI provisions.

Drescher has defended the deal, arguing that it contains the “most progressive AI protections ever written.” Those protections include requirements that studios gain consent and provide compensation for the use of AI to generate digital doubles. If an AI-generated synthetic character has a recognizable facial feature of a real actor and that actor’s name was contained in the prompt to generate that character, the producer must get the okay from the actor. The deal also mandates studios to get permission from the estate of a deceased actor before using AI to reanimate them; if the actor does not have an estate and representatives cannot be found, the producer can get the greenlight from the union. Further, the TA requires that actors be paid at least the minimum day rate for the process of having their faces and bodies scanned.

As for synthetic performers, i.e., assets meant to look and sound like humans but not specifically based on any one real person, studios would have to notify SAG-AFTRA of their intent to produce a project using an AI asset instead of a real person and offer the union an “opportunity to bargain in good faith” to give human actors a chance to audition for those roles.

Still, actor anxiety abounds. The provisions do not prohibit AI, nor stop studios from training the technology on actors’ performances to generate “synthetic” performers. Studios could potentially use derived lookalikes that don’t rise to the level of replicas, and one stipulation allows the bypassing of consent if the likeness “would be protected by the First Amendment (e.g., comment, criticism, scholarship, satire or parody, use in a docudrama, or historical or biographical work).” Plus, performers will be responsible for negotiating their own compensation for contracted use of digital replicas developed without their physical participation — and as any union member knows, individual negotiations with a large employer are always an uphill battle.

Also of concern is studios’ ability to require consent to AI as a condition of employment. Performers fear that this will effectively coerce those most in need of work to offer such assent despite their reservation, a possibility SAG-AFTRA’s national executive director and chief negotiator Duncan Crabtree-Ireland acknowledged in a lengthy membership meeting. Such a condition could further divide the membership, with the highest-profile members holding the leverage to play hardball while those without such standing strong-armed into submitting.

“The reality is, there has never been a time when we have been able to successfully just block technology from advancing,” said Crabtree-Ireland. “And so strategically, our best option is to channel that technology in the best possible direction.”

The Ratification Vote

Despite some members’ objections to the use of any AI, a total — or even near-total — prohibition on the technology was never on the table. According to a document released by SAG-AFTRA when the strike began on July 13, the union proposed “a comprehensive set of provisions to grant informed consent and fair compensation when a ‘digital replica’ is made or our performance is changed using AI.” The vision was always guardrails on AI, not a ban.

That isn’t to dismiss members’ concerns; this is a new technology, one that is advancing far more quickly than a three-year contract, and the matter of requiring consent to AI as a condition of employment stands out as a particularly well-founded problem. And it is true that the introduction of AI not only threatens performers, but other crew members, too. If one has fewer human actors, one needs fewer hair and makeup artists. The latter issue is why Production Assistants United, the nascent production assistants’ (PAs) union, is urging performers to vote against ratification, even as such a vote could extend the period that PAs and other crew members are out of work.

But why were so many SAG-AFTRA members unaware of the substance and scope of the union’s AI proposals? And why have their concerns seemingly taken much of the leadership by surprise? The answer, at least in part, has to do with the nature of SAG-AFTRA’s structure and communications: unlike the WGA, which is particularly democratic and transparent in its communications with membership, the actors’ union falls on the other end of the spectrum.

Rejecting a TA tends to require strong rank-and-file networks, and thanks to the shabby state of SAG-AFTRA internal organizing, it is unclear if the members will pull that off this time. Indeed, the fact that members have so rarely rejected a TA for the TV/Theatricals contract is itself evidence of the need for internal reform.

But the ratification vote is likely to be much closer than the leadership would hope, and responsibility for that lies with those at the top of the union. If members desired blanket prohibitions on AI, the leadership should’ve been aware of that when proposals were formulated. And if members desire a truly fighting union, it must be built from the bottom up, with SAG-AFTRA leadership pushed to boost the transparency of their communications and their responsiveness to the rank and file — not just during a ratification vote, but all the time.