Section 1: AI Filmmaking Pipeline and Responsible Use · Lesson 3 of 5 (3 of 45) · 60 minutes · Evidence: Authorship decision log
Imagine two students hand in nearly identical films. One kept every suggestion the generator gave. The other rejected most of them, rewrote a key scene, and can tell you why. Who made the film? The honest answer is that both used the same tool, but only one of them acted like an author. Today you learn to see the difference, and to build a record that shows it.
Production journal: write your answer below.
What must a filmmaker understand before they can separate student decisions from generated suggestions and document meaningful revision?
A generation tool is a fast, tireless suggestion engine. It can draft a scene, propose a shot, or produce a face. What it cannot do is care what the film means. Authorship is the set of decisions that give the film a point of view, and it is yours only if you actually make those decisions. Not every decision is equal, so we sort them.
No matter how capable the tool, you keep control of these five:
Revision means improving work after seeing it. Surface revision swaps a word for a synonym or nudges a color. Meaningful revision changes what the work says: a character’s motive, the order of reveals, the ending. When you document a revision, state what the draft said before, what it says now, and why you changed it, tied back to your intent.
Provenance is the record of an asset’s origin and history: who or what made it, with which tool, when, under what permission, and what you changed afterward. In the decision log, each row carries a provenance note, so you can later say exactly which words, images and sounds were generated, which you wrote, and which you transformed.
The decision log is a running table. Every time a tool offers you something, log the stage, what was offered, what you did (accepted, rejected or rewritten), and one reason connected to your intent. Rejections count; they are some of your best evidence.
For a comedy short about a robot vacuum that fears the dark, a student’s intent says: “gentle comedy about being scared of small things.” The generator offers a script in which the vacuum becomes a villain. The student rejects it: “tone is menacing, contradicts intent.” The next draft has the right tone but a flat ending, so she rewrites it: the vacuum finally cleans the dark corner, and the family finds a night-light it left behind. Log row: Script v2, ending / generated, rewritten / meaning changed from defeat to courage / provenance: generated draft, final ending written by me.
Watch again how a creator uses an AI-assisted editor, this time focusing on who is in control at each moment. Sometimes the creator gives an instruction and the tool executes it exactly. Sometimes the tool makes a choice nobody asked for, such as picking a visual, a transition or a piece of music. Those unrequested choices are the ones an author must notice.
Sort what you see into three kinds of moments: creator decided, tool suggested and creator accepted, and tool decided and creator may not have noticed. The third kind is where authorship quietly slips away. A filmmaker who never reviews those choices is not controlling the film.
For every decision point, write a short log row of your own as if you were the creator: what was offered, what you would have done, and the reason. You are practicing the habit that the independent studio asks you to keep for the rest of the course.
As you study, record each decision point in your production journal.
| Decision point | What the filmmaker does | What the tool automates | What still needs human review |
|---|---|---|---|
| Supply the story direction | Writes the idea, tone and audience in their own words | Expands the idea into scenes or script text | Whether the expansion still carries the original intent |
| Review the generated visuals | Keeps, rejects or asks for new versions | Picks and matches visuals to lines | Whether images are consistent, appropriate and truly match the meaning of each line |
| Choose music and voice | Selects mood and checks levels against the voice | Offers a library and a default voice | Whether the audio is licensed, allowed, and honest about being synthetic |
| Final preview and approval | Watches the whole piece and signs off or sends it back | Plays and renders the result | Whether any unrequested tool choice changed the meaning |
Studio task: Compare an unedited generation with a revised version and identify the human decisions.
Your teacher will give you two versions of the same short scene: an unedited generation (exactly what a tool produced) and a revised version (the same scene after a student worked on it). For practice, here is a sample pair for a thirty-second science-fiction scene about a botanist on a space station.
Unedited generation: “Dr. Lin walks into the greenhouse. The plants are dying. She feels sad. She says, ‘I must save them.’ She waters the plants. The plants recover. The end.”
Revised version: “Dr. Lin enters the greenhouse. Every leaf has curled except one, which leans toward the door. She says nothing. She hesitates, then switches off the grow-lights and sits in the dark with it.”
Coaching tips. A change that only swaps words is surface. A change that alters what a character wants or what the audience feels is meaning. If you cannot connect a change to an intent, say so; it may be a weak revision. Notice that “she says nothing” is a removal, and removals are decisions too.
Film intent (one sentence, written before generating): ______ # | Stage | What the tool offered | My decision (accepted / rejected / rewritten) | Reason tied to intent | Provenance note 1 | | | | | 2 | | | | | 3 | | | | | Meaning change (before > after > why): ______
Studio task: Begin a decision log that records accepted rejected and rewritten material.
Begin the decision log you will keep for the whole course. You are not making the film today; you are building the habit and the format.
If your text generation tool is unavailable, write two drafts of the logline yourself, treat one as the “offered” draft, and log the same way. The skill is the log, not the tool.
Checklist before you submit:
Describe one revision that changed meaning rather than surface wording.
Deliverable: Authorship decision log.
Submit the authorship decision log. The file must use the course naming convention and include enough process evidence to show the student's decisions and revisions.
File name (course naming convention): AIF_S01-L03_authorship-decision-log_v01_YOURINITIALS
| Criterion | Advanced | Proficient | Developing | Beginning |
|---|---|---|---|---|
| The work addresses the stated story or production purpose. | Log is anchored in a clear written intent and every row serves the film’s stated purpose. | Log is anchored in a written intent and entries relate to the film’s purpose. | Intent is vague or entries are loosely related to it. | No intent or entries unrelated to the film. |
| The student applies the required method accurately and produces usable capstone evidence. | Six or more rows with decision, reason and provenance, and the three decision types used correctly. | Six rows with decision, reason and provenance mostly complete. | Fewer rows, or reasons and provenance often missing. | Log is a list of tool outputs with no decisions. |
| The student identifies at least one quality issue and records a meaningful revision decision. | Meaning revision clearly shows before, after and why, and changes what the work says. | Revision shows a real change and a stated reason. | Revision is mostly surface wording. | No revision recorded. |
| The work complies with consent rights attribution privacy disclosure and school policy. | Authorship statement is precise and honest about generated versus human work; sources and permissions are logged. | Authorship statement is honest and sources are logged. | Statement is vague about what was generated. | Presents generated work as wholly human-made or logs no sources. |
Requirement: Students must not present generated output as wholly human-created work.
Honesty about how your film was made is part of its quality. Never present generated output as wholly human-created work. Use specific language such as “script drafted with a text generation tool and rewritten by me” instead of “written by me.” Your decision log and its provenance notes are what let you do that accurately. If you are ever unsure whether something counts as generated, treat it as generated and say so; your teacher would rather read a precise statement than discover a vague one.