When Academy Award-nominated actor Colman Domingo voted "yes" on the 2023 SAG-AFTRA contract that ended Hollywood's historic 118-day strike, he became one of the most prominent voices supporting what many consider the entertainment industry's first comprehensive attempt to regulate artificial intelligence. Yet behind his carefully measured endorsement lies a cautionary tale about how machine learning is fundamentally reshaping not just acting, but the very concept of human performance itself.
This guide explores how Colman Domingo's involvement in the 2023 SAG-AFTRA negotiations illuminates the broader collision between AI & Machine Learning and creative industries. You'll discover why AI-generated digital replicas represent both an existential threat and potential opportunity for performers, how deepfake fraud is expected to reach $40 billion by 2027, and what the entertainment industry's regulatory battles reveal about AI governance challenges facing every sector. Whether you're a technologist tracking AI applications, a business leader concerned about digital identity protection, or simply curious about the future of human creativity in an AI-driven world, this analysis reveals why the Colman Domingo approach to AI negotiation matters far beyond Hollywood.
In December 2023, as SAG-AFTRA members debated whether to ratify a contract that would fundamentally alter how artificial intelligence could use their likenesses, Colman Domingo—then receiving Oscar buzz for his performance in "Rustin"—took a clear stance. "I believe that we have an incredible deal, I believe it's thoughtful and it's about moving the needle forward," Domingo told the Associated Press, adding "I'm very happy with it. I voted yes."
This wasn't a casual endorsement. The AI provisions in the 2023 contract represented Hollywood's first major attempt to establish guardrails around machine learning applications that could digitally replicate actors' voices, faces, and performances. The strike was led by changes in the industry caused by streaming and its effect on residuals, as well as other new technologies such as AI and digital recreation.
The stakes were profound. Duncan Crabtree-Ireland, the union's chief negotiator, warned that studios wanted to scan background actors and then use artificial intelligence to place those actors in other projects "for the rest of eternity" without consent. What made Domingo's position particularly noteworthy was his willingness to support compromise in an environment where some prominent actors—including former SAG-AFTRA board member Justine Bateman—argued the union should have blocked AI usage entirely.
The contract Domingo supported distinguishes between two critical AI-generated categories. "Digital Replicas" are replicas of a specific actor's voice and/or likeness created using digital technology (including generative AI), while "Synthetic Performers" are characters created by digital technology that appear to be natural performers but are not recognizable as any identifiable performers.
For Digital Replicas based on actual performers like Colman Domingo, the contract establishes that creating an Employment-Based Digital Replica requires producers to give actors 48 hours notice, obtain "clear and conspicuous" consent from actors to use their likeness, and pay them for it. This seemingly straightforward protection masks deeper complexities that have troubled critics.
The protections include significant loopholes: for AI involving the participation of human actors, Schedule F performers who make more than $80,000 for a film are exempted, and consent is not required from actors if the photography or soundtrack remains "substantially as scripted, performed, and/or recorded."
The implications extend far beyond contract negotiations. Machine learning systems are increasingly analyzing performances by actors like Colman Domingo to create more sophisticated AI-generated content. Domingo's consistent delivery of powerful and memorable performances serves as a rich source of inspiration for AI-driven creative processes; for example, training AI models on the subtle emotional arcs he portrays in films like "Rustin" or "The Color Purple" can lead to more sophisticated and authentic AI-generated character moments.
This creates a fascinating paradox: the very excellence that makes Domingo's performances valuable also makes them ideal training data for systems that could eventually reduce demand for human actors. The nuanced emotional intelligence and detailed character work that Domingo brings to his film roles serve as a vital benchmark for AI-driven creative tools, with platforms like ReelMind.ai using advanced AI video generation capabilities and sophisticated models to analyze and learn from such depth of performance.
Modern deepfake technology relies on Generative Adversarial Networks (GANs) and advanced neural networks that can analyze thousands of data points from an actor's previous work. Martin Scorsese's "The Irishman" used cutting-edge CG trickery to de-age its stars, Robert De Niro, Al Pacino and Joe Pesci, employing tools like FaceSwap which uses machine learning to comb through countless reference images of actors in their earlier movies.
The technology has evolved rapidly. "In the last year [deepfakes have] gotten so good that even the PhDs on my team with their eyes can't tell the difference," said Ben Colman, CEO of Reality Defender, a deepfake detection company. This deterioration in human detection capability has massive implications for fraud prevention, identity verification, and creative authenticity.
While Hollywood debates artistic and employment implications, AI-generated content has already spawned a massive fraud ecosystem. Generative AI fraud in the U.S. alone is expected to hit $40 billion by 2027, according to the Deloitte Center for Financial Services, with fraud losses climbing from $12.3 billion in 2023 to $40 billion by 2027—a compound annual growth rate of 32%.
These aren't hypothetical projections. Deloitte's Center for Financial Services documented deepfake fraud losses of $12.3 billion in 2023 and predicts they could climb to $40 billion in the U.S. by 2027. The acceleration is staggering: fraud attempts spiked 3,000% in 2023, with 1,740% growth in North America.
The celebrity impersonation problem mirrors what performers like Domingo face in contract negotiations. In early 2025 alone, celebrities were targeted 47 times by AI-generated impersonations, an 81 percent increase compared to all of 2024. Voice cloning has become particularly pernicious, requiring as little as 20-30 seconds of audio to create convincing replicas.
The idea of deepfakes goes back at least a decade, starting within entertainment with special effects requiring high-powered computers and significant time; what's really happened over the last few years has been the democratization of both the tools—now available to anybody with a search—and the democratization of the technology needed to run the software.
This accessibility transforms deepfakes from a specialized Hollywood concern into a universal threat. "Similar to computer viruses, good deepfakes can be incredibly difficult to detect, but unlike computer viruses, good deepfakes can be created by anyone—for example, any junior high school student can create a deepfake using a 5 year old iPhone," says Ben Colman from Reality Defender.
What can other industries learn from how Colman Domingo and SAG-AFTRA approached AI regulation? The entertainment industry's negotiations offer a template—albeit imperfect—for balancing technological innovation with human rights protection.
The core principle that Domingo supported centers on informed consent. SAG-AFTRA wants informed consent—which would be bargained separately at the time of use, when performers would actually know what they're consenting to, rather than boilerplate consent buried in initial contracts.
This approach recognizes that meaningful consent requires:
Crucially, Domingo supported a contract that treats AI regulation as an evolving process rather than a one-time settlement. SAG-AFTRA secured a promise to meet "regularly" with producers to discuss potential pay for performances being used to train a generative AI system, with chief negotiator Duncan Crabtree-Ireland acknowledging: "What I expect to happen over the next couple of years, as these terms play out, is that we'll be getting feedback from our members on what's working, what needs to be improved upon."
Implement multi-factor verification beyond visual/audio: Given that human detection of deepfakes has dropped to 24.5% accuracy for high-quality video, organizations should establish procedural safeguards—code words, callback protocols, and out-of-band verification—rather than relying on employees' ability to spot fakes. The Colman Domingo contract model emphasizes systemic protections over individual vigilance.
Structure AI consent as project-specific, not perpetual: Following the SAG-AFTRA framework that Domingo supported, any agreement involving AI use of personal data, likeness, or creative work should specify exact applications with time limitations. Avoid blanket authorizations that allow unlimited future use—insist on separate negotiation for each new application, ensuring you maintain control as technology capabilities expand.
Build regular AI audit mechanisms into agreements: The best Colman Domingo guide principle is treating AI governance as an ongoing process. Whether you're negotiating employment contracts, vendor agreements, or industry standards, include mandatory periodic reviews (semi-annual or quarterly) to assess how AI applications are actually being deployed versus initial agreements, with rights to renegotiate as circumstances change.
Q: What specific AI protections did Colman Domingo vote for in the 2023 SAG-AFTRA contract?
A: Domingo supported provisions requiring 48-hour notice before creating digital replicas, "clear and conspicuous" consent from actors for AI use of their likenesses, and compensation for digital replica creation and usage. The contract distinguishes between Employment-Based Digital Replicas (created during a performer's work) and Independently Created Digital Replicas, with different protections for each. However, the agreement includes exemptions for high-earning performers (Schedule F) and doesn't require consent when footage remains "substantially as scripted."
Q: How does machine learning use performances by actors like Colman Domingo to train AI systems?
A: AI platforms analyze thousands of data points from an actor's previous performances—facial expressions, vocal patterns, emotional arcs, movement dynamics—to train models that can generate similar content. Training on Domingo's nuanced performances in films like "Rustin" helps AI systems learn to create more sophisticated and emotionally authentic digital characters. This creates a paradox: the qualities that make master performers valuable also make them ideal training data for systems that could reduce demand for human actors.
Q: Why are deepfake fraud losses expected to reach $40 billion by 2027?
A: The explosive growth stems from three factors: democratization of deepfake creation tools (now accessible via simple apps rather than requiring specialized expertise), the rapid improvement in quality making deepfakes nearly undetectable to humans, and the 3,000% spike in fraud attempts as criminals recognize the technology's effectiveness. Voice cloning—requiring only 20-30 seconds of audio—has become particularly prevalent, with North America experiencing a 1,740% increase in deepfake fraud between 2022 and 2023. The Deloitte projection of $40 billion by 2027 (up from $12.3 billion in 2023) represents a 32% compound annual growth rate.
Q: What can other industries learn from the Colman Domingo approach to AI negotiations?
A: The key lesson is that effective AI governance requires ongoing negotiation rather than one-time agreements. Domingo supported a contract that includes regular meetings between unions and producers to discuss AI applications, semi-annual reviews of generative AI use, and the right to renegotiate as technology evolves. This framework acknowledges that AI capabilities will continue advancing in unpredictable ways, making static regulations obsolete. Industries should adopt similar iterative approaches with specific consent requirements, project-by-project authorization (not blanket permissions), and built-in review mechanisms to adapt protections as AI applications expand.
When Colman Domingo cast his vote supporting the 2023 SAG-AFTRA contract, he wasn't just making a decision about his own career—he was helping establish the first serious framework for how human creativity and artificial intelligence might coexist. The contract's limitations and loopholes have drawn legitimate criticism, but its core principle represents a crucial stake in the ground: AI cannot simply appropriate human identity, performance, and creativity without consent and compensation.
The $40 billion in projected deepfake fraud losses by 2027 demonstrates that these questions extend far beyond Hollywood. Every organization now faces the challenge of protecting digital identity in an era when machine learning can convincingly replicate voices from 20 seconds of audio and create photorealistic video from still images. The democratization of deepfake technology means that the same tools Hollywood uses for de-aging effects are now available to anyone with a smartphone and internet connection.
Yet perhaps the most profound insight from the Domingo approach is recognizing that AI governance cannot be a one-time settlement. Technology evolves too rapidly; capabilities that seem impossible today become commonplace within months. The iterative framework—regular reviews, ongoing negotiation, the right to renegotiate as circumstances change—offers the only realistic path forward.
As you consider AI's implications for your own industry, ask yourself: Are you building static defenses against a moving target, or creating adaptive frameworks that can evolve alongside the technology? The best Colman Domingo guide isn't just about the specific protections he supported—it's about the principle that human dignity and technological progress need not be opposing forces, if we're willing to negotiate the terms of their coexistence thoughtfully and continuously.
What framework is your organization putting in place to ensure AI serves human flourishing rather than replacing it?
Related Free Tool
Readability Checker
Measure your content's Flesch Reading Ease score instantly.
Get the latest AI-powered insights delivered to your inbox every week. No spam, ever.
Unsubscribe anytime. We respect your privacy.
Written by
Marcus ReidHealth & Science
Health and science writer dedicated to translating complex medical and scientific research into accessible, actionable insights.
Loading comments...
Read Next
Health & ScienceSleep deprivation costs the U.S. $411 billion annually. Discover the neuroscience behind rest and evidence-based strategies to sleep better starting tonight.
Lance Stroll's F1 career intersects with AI revolution as Aston Martin harnesses 1.1 million data points per second to transform racing through machine learning.
The viral 82-0 basketball game isn't just entertainment—it's a masterclass in ML algorithms, era-adjusted data normalization, and optimization theory.
A well-designed prompt can boost AI accuracy by 58%. Learn the basics of prompt engineering to unlock superior AI performance without expensive upgrades.