When an AI-Made Pitch Film Hurts Your Project

Our infrastructure includes generative tools. Saying that plainly is important before the rest of this, because what follows would otherwise read as marketing from someone with a different product to sell. We do not have a different product. We are describing the cases where our own work is the wrong choice.

One: when the project's route runs through a body with restrictions

Festivals, funds and some institutional financiers have been introducing rules about generatively produced work, and the rules differ between them and have been changing quickly. Some require declaration. Some restrict eligibility in specific sections. Some are silent, which is not the same as permissive.

If your project's plan depends on a specific body, the method is not a neutral production decision. It can affect a route you are counting on. The answer is not to guess, and it is certainly not to take our word for it. Read the current regulations of the specific bodies in your plan, this year, before anyone starts.

We never promise eligibility anywhere, for anything. Nobody can, because the rules are not ours and they change.

Two: when the film has to prove that you can direct

A significant share of proof of concept films exist to answer a question about a person rather than about a project: can this director handle this material at this scale.

If that is the question, then how the film was made is part of the evidence, and a film produced in a pipeline the director does not control is weaker evidence, sometimes to the point of being counterproductive. The room may admire it and still not have learned the thing they needed to learn.

This is fixable, but only by structuring the work so the director actually directs it rather than approves it. That is a different engagement, and it needs to be agreed at the start.

Three: when it involves a real person

Generating a recognisable human being, living or recently dead, is a legal and ethical problem before it is a creative one. Rights of publicity and likeness vary by jurisdiction and are actively being litigated and legislated.

There is also a practical trap that catches people. Generating without any reference does not make you safe. Models produce faces that can strongly resemble real performers, and a resemblance that is accidental is still a resemblance in the eyes of the person who has it. If a specific individual appears, it should be with their agreement, on a documented basis.

Four: when the audience for the film is the wrong audience for the method

Reactions differ sharply by room. A founder raising capital, a developer selling an unbuilt project, a brand launching a product: in these contexts the method is usually irrelevant, and occasionally an advantage, because it demonstrates speed and control.

In parts of the film industry the reaction is different, and it is not uniform. Some people care a great deal, some do not care at all, and you often cannot tell which you are dealing with in advance. If your route runs through a small number of specific people whose position on this you do not know, that is worth finding out before you commission, and it is a cheap thing to ask.

Five: when the project is about the very thing the method undercuts

Occasionally a project's central claim is about human craft, physical performance, a documentary relationship with reality, or a place captured rather than constructed. In those cases the method can quietly contradict the pitch, and viewers feel the contradiction before they can articulate it.

This is a taste judgment rather than a rule, and it is one worth making consciously instead of discovering in a room.

What we actually do about this

• We ask, in the first conversation, what route the project is taking and who has to say yes.
• We say plainly that we do not promise eligibility with any festival, fund or institution, and we put that in the scope of work.
• We do not put a specific real person on screen without a documented basis.
• Where the question is whether you can direct, we structure the work so that you direct it.
• If we think the method is wrong for the project, we say so, and we would rather lose the job than deliver a film that damages the thing it was meant to help.

The reason we publish this

Partly because it is true. Also because the alternative is worse for us. A film that harms a project produces a client who regrets working with us, and that is more expensive than any single job is worth.

The tools are infrastructure. They are not a reason to hire anyone, and they are not a reason to hire us. Direction is.

How we work, including what we refuse, is on the proof of concept page.

There are five situations where a generatively produced pitch film can damage a project: restricted routes, films that must prove a director, real people on screen, rooms where the method is contested, and projects whose claim the method contradicts. Verify the rules at the source and never accept a promise of eligibility from anyone.