‘Digging the grave of my profession’: the Hollywood creatives training AI to do their jobs
Sunday, 23 August 2026
Warm-up
- Where do you see AI already changing film or TV work?
- Would you take short-term work that might replace your job? Why?
- Which production task would you find hardest to teach to a machine?
Vocabulary
- lucrative
- bringing in a lot of money compared to the time or effort
- replicate
- to copy a result or skill so it works in the same way
- hypothetical
- imagined for practice or testing, not a real case
- schedule
- a plan that shows what happens and when during a project
- permit
- an official document giving legal permission to do something
- transcribe
- to write down exactly what is said in audio or video
- freelance
- working for yourself and taking short jobs for different clients
- fatalistic
- believing events cannot be changed, so they must be accepted
Reading
In Hollywood, some creatives are taking lucrative gigs to train AI models to replicate their work, pushed by a jobs slump. Trainers earn $12 to $200 an hour via agencies serving Anthropic and OpenAI. Feelings range from realism to a fatalistic sense of loss; one worker said he was "handed a shovel and asked to dig the grave of my profession".
Ruth Fowler, a Los Angeles screenwriter and producer, now teaches systems how production decisions are made. She built a hypothetical two-day shoot and showed an AI how to create a schedule from emails, a script and materials, including permits, hazard checks, daylight needs, crew lists and child protection. She also trained an AI to assemble a pitch deck for film and TV ideas. It is production work these systems are being taught to imitate.
An LA documentary director trained a model to transcribe speech from a Little League game despite loud music and crowd noise; he mapped speakers and noted accents, which worried some freelance staff.
Industry data show the pressure: Netflix used AI in 300 of its 1,000 titles in 2026, and director Ron Howard is embracing the technology. Shoot days in Los Angeles fell 48% between 2021 and 2025, FilmLA says.
Comprehension
- Why are some Hollywood creatives training AI now?
- How much are trainers paid per hour for this work?
- Which AI companies are named as the agencies’ clients?
- What did Ruth Fowler teach an AI to produce for a shoot?
- By how much did LA shoot days fall between 2021 and 2025?
Grammar focus
Adding emphasis - cleft sentences
Use a cleft (It + be + X + that/who) to emphasise or focus one part of a sentence. It puts attention on X. (Examples constructed.)
- It was **the crew that** mapped the speakers at the game.
- It is **production work that** these systems are taught to imitate.
Grammar exercise
Circle the correct option in each sentence.
- Many trainers said it was pay (that / what) drew them in.
- Some worry it was background noise (what / that) ruined the audio.
- Producers claimed it was production work (that / what) AI imitates.
- Many agreed it was freelancers (that / who) felt threatened.
- Industry figures said it was Netflix (that / what) used AI in titles.
Discussion
- Is it ethical to train AI that could replace your role? Why or why not?
- Which film or TV tasks should never be automated? Explain your choice.
- Who should control creative data used to train AI tools?
- Do Netflix’s 300 AI-linked titles show progress or risk for artists?
Vocabulary, reading, and exercises with instant feedback