A new tool changes what is easy to try. It can also change who gets paid, what a team trusts and where responsibility becomes harder to see. Both belong in a conversation about AI and filmmaking.
Some assurances about work and creativity are too certain. AI can assist with parts of production; nobody can promise that its adoption will protect every role or make the industry fairer. A more useful question is how to use it with judgement and care.
Start with the task, not the tool
A transcript to search. Research notes to organise. A visual idea to test before committing to a shoot. These are bounded needs against which a team can judge whether an assistant helps.
Decide what a good result would look like and what it would take to check it. Include correction time, subscriptions, source permissions and the effect on collaborators. A quick first output can become an expensive process if nobody can establish where it came from or whether it is usable.
An experiment should give you information, even when the answer is that an ordinary conversation or existing editing tool does the job better.
Keep meaning and evidence separate
Generated material can look convincing without documenting something that happened. That distinction matters when the audience expects a film to show a real person, place or event.
Do not pass a generated historical image off as archive. Do not use a synthetic voice or likeness without the appropriate permission. Keep a record of what was made, where a source came from and how the material is described to viewers.
The original essay raised concerns about the historical record. Those concerns remain part of the creative decision, not a small disclaimer to add after the edit. If a sequence changes what an audience believes is evidence, its treatment needs deliberate thought.
Human judgement is not a ceremonial last click.
Access is more than an available tool
An independent filmmaker may be able to test an image or workflow that previously required resources they did not have. That is a meaningful possibility. It is not the same as equal access to making a film.
People also need time, training, equipment, money and somebody to ask when a tool fails. Paid tiers and uncertain terms can add barriers. Environmental and resource costs should be considered too; a generated image is not immaterial because its production is out of sight.
If access matters, build support around the tool. Share a method that somebody can inspect, adapt and question. Give them a route to begin without requiring a particular subscription.
Treat collaborators as part of the decision
“Efficiency” can describe saved effort or somebody else’s lost income. Talk about which tasks are changing, who is checking the result and how credit and payment will work. Do not assume that a team shares the same appetite for experimentation.
Human judgement is not a ceremonial last click. Performance, rhythm, meaning and context are the work. A suggested cut needs to be watched. A research summary needs its sources checked. A striking image still needs to belong in the film.
Make room to learn without promising the future
At JBN, we are building and sharing tools alongside production. Our workshops, workflow clinics and partner-funded pilots are ways to investigate practical needs with the people doing the work.
The aim is more agency: understand what a tool can do, notice where it fails and decide whether it earns a place in your process. That is a better foundation than either automatic enthusiasm or a promise that nothing will change.
