New SAG-AFTRA Game Localization Contract Restricts AI Usage in Dubbing

On November 14, 2024, the Screen Actors Guild-American Federation of Television and Radio Artists (SAG-AFTRA) announced an updated version of a previous agreement that covers the localization of video game projects produced in a non-English language.

With approximately 160,000 members, SAG-AFTRA calls itself the “world’s largest union representing performers and broadcasters.” The union also represents voiceover artists, including those who provide dubbing. 

As comics and gaming website Bleeding Cool reported, the new Independent Interactive Localization Agreement is essentially an updated version of the base terms from the union’s Tiered Budget Independent Interactive Media Agreement, plus AI protections.

The new agreement is signed on a project-by-project basis by employers whose project was originally scripted in a language other than English, and whose intellectual property owner is based outside of the United States.

“Many brilliant, beloved games come to market in the U.S. from other countries, projects which need highly skilled localizing performers,” Interactive Media Agreement Negotiating Committee Chair Sarah Elmaleh was quoted as saying in the press release. Elmaleh added that “[m]any such companies have already signed Interim Localization Agreements”. 

The contract was reportedly crafted “based on direct feedback from the community that does this work.”

However, the introduction of the Independent Interactive Localization Agreement does not interrupt an ongoing SAG-AFTRA strike.

The union called for the strike, effective July 26, 2024, in response to stalled negotiations, which began in October 2022. The use of AI still presents a major hurdle, as SAG-AFTRA explained:

“Although agreements have been reached on many issues important to SAG-AFTRA members, the employers refuse to plainly affirm, in clear and enforceable language, that they will protect all performers covered by this contract in their A.I. language.”

The strike applies to over 130 video game projects currently signed to the union’s interim and independent agreements; the goal is for those signatories to sign updated agreements. 

Human-Made Recordings and Digital Replicas

The Interactive Media Agreement, originally introduced in 2017, was extended until 2022. At that point, it was replaced by the Tiered-Budget Independent Interactive Media Agreement along with the Interim Interactive Media Agreement and, most recently, the Independent Interactive Localization Agreement.

Katie Sikkema, a union contracts consultant, explained on LinkedIn that the updated agreement “includes a mandatory buyout of all reuse and integration for an additional 50% of scale, which means the IP owner can use the (original, human-made) recordings for whatever they want with no further fees due (except merchandising/talking toys, which must be separately negotiated).”

The contract describes generative AI (GenAI) as a “subset of AI that learns patterns from data and then produces content based on those patterns.” 

Interestingly, GenAI is considered separate from “digital replicas” of performers, including voice actors, though the term is not defined. To that point, employers are required to provide the union advance notice if they intend to use GenAI to generate material other than digital replicas. 

Employers that want to create GenAI material using prompts including a performer’s name, or a unique character associated with that performer, must get consent from the performer and bargain for the use of GenAI material at a specific minimum rate.

SAG-AFTRA’s strikes have brought attention to the issues GenAI presents for the entertainment industry, an area ripe with opportunities for language AI startups specializing in dubbing. In March 2024, the union ratified the 2023 Television Animation Agreement, which includes “strong protections around the use of artificial intelligence” for voice acting and other performances. 

Google Says There’s a Better Way to Create High-Quality Training Data for AI Translation

In an October 14, 2024 paper, Google researchers highlighted the potential of AI translations refined by humans or human translations refined by large language models (LLMs) as alternatives to traditional human-only references.

Talking to Slator, Zhongtao Liu, a Software Engineer at Google, explained that their study addresses a growing challenge in the translation industry: scaling the collection of high-quality data needed for fine-tuning and testing machine translation (MT) systems. 

With translation demand expanding across multiple languages, domains, and use cases, traditional methods that rely solely on human translators have become increasingly expensive, time-consuming, and hard to scale.

To address this challenge, the researchers explored more efficient approaches to collect high-quality translation data. They compared 11 different approaches — including human-only, machine-only, and hybrid methods — to determine the most effective and cost-efficient one.

Human-only workflows involved either a single human translation step or included an additional one or two human review steps. Machine-only workflows ranged from single-step AI translations using top AI systems — MT systems or LLMs — to more complex workflows, where AI translations were refined by an LLM. Hybrid workflows combined human expertise and AI efficiency; in some cases, AI translations were refined by humans (i.e., post-editors), while in others, human translations were refined by LLMs.

They found that combining human expertise and AI efficiency can achieve translation quality comparable to, or even better than, traditional human-only workflows — all while significantly reducing costs. “Our findings demonstrate that human-machine collaboration can match or even exceed human-only translation quality while being more cost-efficient,” the researchers said.

The best combination of quality and cost appears to be human post-editing of AI translations. This approach delivered top-tier quality at only 60% of the cost of traditional human-only methods, while maintaining the same level of quality.

“This indicates that human-machine collaboration can be a faster, more cost-efficient alternative to traditional collection of translations from humans, optimizing both quality and resource allocation by leveraging the strengths of both humans and machines,” they noted.

The researchers emphasized that the quality improvements stem from the complementary strengths of human and AI collaboration, rather than from the superior capability of either the AI or the human (post-editor) alone, underscoring the importance of leveraging both human and AI strengths to achieve optimal translation quality.

They noted that LLMs were less effective than human post-editors at identifying and correcting errors in AI-generated translations. On the other hand, human reviewers tended to make fewer changes when reviewing human-generated translations, possibly overlooking certain errors. Interestingly, even additional rounds of human review did not substantially improve the quality. This observation supports the argument for human-machine collaboration, where each component helps address the other’s blind spots, according to the researchers.

“These findings highlight the complementary strengths of human and machine post-editing methods, indicating that a hybrid method is likely the most effective strategy,” they said.

Authors: Zhongtao Liu, Parker Riley, Daniel Deutsch, Alison Lui, Mengmeng Niu, Apu Shah, and Markus Freitag