The popular story goes like this: the people who know AI best are terrified of it, so the rest of us should be too. Hollywood’s labor unions have a different read. According to The Verge AI, SAG-AFTRA and the Writers Guild of America East responded to the latest round of extinction warnings from AI researchers with a pointed message: stop looking at the horizon and look at what’s already on the ground.
The Verge AI also asked Disney, Netflix, Amazon, Lionsgate, and several AI film startups for their take on the warnings. None replied. The unions did, and their answer says a lot about who actually has skin in this game.
The warnings and the skeptics
The backdrop, as The Verge AI details it, is a wave of alarm from inside the frontier labs. Former Anthropic researcher Jacob Coxon quit to sound the alarm. Anthropic alignment lead Evan Hubinger said the technology could “kill all humans” at some point “within the next decade.” Heads of Anthropic, OpenAI, Google DeepMind, and X have called for an industry-wide slowdown. President Trump has dismissed all of it as overreaction, while some other US politicians feel pressure to legislate.
SAG-AFTRA chief negotiator Duncan Crabtree-Ireland isn’t dismissing the risk, but he’s not taking the messengers at face value either. He told The Verge AI that hearing doom predictions from “the CEOs of these companies or the billionaires who’ve made their billions off of doing the exact thing that they’re now warning us might not be good” has left him unsure “what to trust or what to believe.” He suspects some of the extinction talk is hyperbole, and notes that doomerism has become a reliable way for AI companies to generate buzz.
WGAE executive director Sam Wheeler went further. He described “the tech oligarchs pushing AI on our society” as reckless, adding that their recklessness “has been matched only by their rampant dishonesty.”
Why the unions think they were right first
What stands out here is the timing argument. Hollywood’s unions were fighting AI in 2023, before whistleblowing became a trend. WGA members spent 148 days on strike. SAG-AFTRA struck too. Both came away with contract language that requires studios to get permission and pay actors when they use AI-generated replicas.
Crabtree-Ireland’s view is that the contract clause changed the economics. “If that principle hadn’t been established, I think the economic incentive to use replication and to replace human performances would have been much stronger,” he said, and adoption of digital replicas would have been “much greater and faster.”
That’s the contrarian point in a nutshell. The unions didn’t stop AI in Hollywood. Studios still use it. But they built an enforcement mechanism that works today, on real harms, while the frontier labs debate hypothetical ones.
The gap the unions can’t close alone
There’s a limit, and both unions know it. Crabtree-Ireland was clear that the guardrails cover the major employers in entertainment, not OpenAI, Anthropic, or Google directly. Wheeler put it bluntly: “collective bargaining is not a replacement for government regulation,” and he thanked states like New York for acting “in the absence of federal action.”
So the actual picture is layered:
- Labor contracts handle the near-term harm (replicas, replacement, credit)
- State laws are filling in where Washington hasn’t
- The extinction debate runs on a separate track, mostly among the people building the models
My take: those tracks are going to merge. Coxon’s exit got traction with the public, Crabtree-Ireland admitted. Public attention is the lever the unions used in 2023. If the doom narrative pulls voters in, the beneficiaries may be the workers who already have concrete asks on the table.
What this means if you build or deploy AI
A few practical takeaways:
- Consent and compensation for likeness are now a bargained standard in US entertainment. Any product touching voices, faces, or performances should assume that norm spreads to other sectors.
- State-level rules are the near-term regulatory reality. Watch New York and California, not just Congress.
- Credibility matters. The unions’ skepticism toward lab CEOs is a warning to anyone selling AI while also warning about it. Mixed messages erode trust with the exact groups you’ll need to negotiate with later.
- Focused messaging wins. Crabtree-Ireland credited the strike’s success to a clear public explanation of what was at stake and what could fix it. That’s a playbook for any organization trying to shape AI policy.
The studios’ silence is its own data point. The people with the most to gain from AI in production chose not to weigh in on whether it might end the world. The people with the most to lose from it answered within days. Watch where the next round of legislation lands; the full interviews are at The Verge AI.