The robots
went into post

Everyone is arguing about the AI actor on the red carpet. The real shift happened backstage. On 16 July Netflix told investors that generative AI touched around 300 of its films and shows this year, most of it in post. The same week a UK report showed the post and VFX crews are the ones being cut. The robots did not come for the poster. They came for the edit.

01The number Netflix finally said out loud

Netflix has spent two years being careful with the word AI. Last week it stopped hedging. In its second quarter shareholder letter, published 16 July, the company said generative AI was used in around 300 of its films and TV titles this year, spread across concept development, pre-visualisation, post-production and release.

In its own words: it is leveraging these tools to deliver higher quality output more quickly and at a lower cost than traditional methods, and in some cases productions would have had to leave out key shots and sequences without them. That is not a lab demo. That is a company that took 12.6 billion dollars in a quarter, up 13 percent, describing how its films now get made.

02It is landing in post, not on the poster

Look at where the 300 actually live. Co-chief executive Ted Sarandos pointed at one title, The American Experiment, where 17 minutes of finished footage came together twice as fast and at half the cost of the old way. Not a synthetic actor. Not a fake performance. The unglamorous middle of the pipeline: the grade, the clean-up, the plate, the shot that used to get cut in a budget meeting.

Sarandos was blunt about the limit. It takes great artists to make something great, he said, and AI is not changing that. Movies are made by people who make movies. The technology just hands them better tools. The loud fight, the AI actor, the fully generated feature hunting for a buyer at Cannes, is front of house theatre. The actual adoption is quiet, technical and already normal. Netflix even bought a company for it, snapping up the AI outfit InterPositive in March.

03The same week, the bill arrived

On 15 July, one day before Netflix filed, UK training body ScreenSkills published its own read of the industry. The volume of British films and shows released fell 13 percent between 2022 and 2025. The workforce is running at about 50 to 55 percent of capacity, down from 60 to 65 percent two years ago. Good crews are sitting at home.

Now the part that lines up too neatly. The workers hit hardest are in VFX and post-production, where capacity dropped from 45 percent in 2023 to 30 percent in 2025. The same report forecasts AI could touch 80 percent of production roles by 2028, and names animation, post-production and VFX as the most exposed. Netflix is celebrating faster, cheaper post in the exact corner of the pipeline where the crews are already thinnest. The efficiency in one report and the lost hours in the other are the same shots.

04What we'd tell you

Do not pick a side in the shouting match. Learn where the thing actually lands. If you make video, the AI question stopped being will a model replace me. It is now which parts of my pipeline just became table stakes, and which parts are still mine.

Point it at the shots you could not otherwise afford. Let it do the ninth version of a background, the tidy-up, the pre-vis nobody was going to pay a human to render. Then guard the rest with your life. Taste, story, the judgement of what is even worth making. A model can give you a cleaner plate. It cannot decide the film is good.

Netflix said the honest case out loud: AI let productions keep shots they would have thrown away. The honest warning sits one report away, in a crew working at 30 percent. Read both, then aim the tool like you mean it. The robots went into post. What they do there is still up to the people in the room.