Three years after strike, Hollywood implements new protections against AI misuse

Three years after the strike that paralyzed Hollywood, studios and unions have found common ground over concerns about the use of artificial intelligence in film during wage deal negotiations, says chief artist negotiator Duncan Crabtree-Ireland.

Although negotiations on this agreement – which will come into effect in July if approved by union members – have been “very intense”the tone of the discussions was revealed “much more collaborative”assured the manager during an interview with AFP. Duncan Crabtree-Ireland is the chief negotiator of the very influential American actors’ union, which has more than 160,000 members from cinema, television and video games. “I really think that the 2023 strikes, although they were difficult for all of us, helped to revive the relationship between the studios and the unions in general”he adds.

Skip the ad

The American screenwriters, joined by the actors, went on strike to demand better remuneration and supervision of AI. This collective action paralyzed the production of American films and series for several months. Studios and platforms now say they better understand the importance of AI to union members, and agree that most of the work should be done by a human.

Protection against AI

But AI creates “still a lot of worries”says Duncan Crabtree-Ireland. This can be explained by the technological advances made over the past three years, he says, citing the example of the AI-generated character Tilly Norwood. Created by the company Particle6, it caused a sensation when it debuted in a short film broadcast on social networks in 2025.

The new agreement does not close the door on AI, but puts in place new protections against the use of certain forms of this technology, which Duncan Crabtree-Ireland classifies into two categories. First, digital replicas where AI is used “to reproduce a living or past artist”. This technology must have “consent and fair remuneration”specifies Duncan Crabtree-Ireland. Then, the synthetic replicas which include characters created by AI “not based on a real person who exists in the world, but on the information on which the AI ​​was trained”he describes. In this case, the technology does not represent a specific person so there is no need for consent or remuneration.”

The agreement stipulates that synthetic replicas should not be “used only in exceptional cases or unusual circumstances”. Without going as far as a total ban, the agreement creates conditions “very dissuasive” to their recourse. Another protection measure concerns dubbing, to prevent the voice of a main actor in a film from being reproduced in other languages ​​without his consent.