Copyright Office Part 2: prompts alone, and arranging AI video

In Part 2 of its AI report, the US Copyright Office concludes that, based on how current generally available technology works, prompts do not alone provide sufficient control over the expressive elements of an output. It also says authors keep copyright in their own expression that is perceptible in AI outputs, and in the creative selection, coordination or arrangement of material, or creative modifications of the outputs.

This is a reading of a report summary, not legal advice, and the Office says sufficiency of human contribution is decided case by case. Read 2026-10-01.

What does the report conclude?

The report's conclusions, as summarised in its introduction, are in the table.

Which Sume edit routes are arrangement steps?

Three Sume routes change or arrange material rather than generate it. Video trim returns a new MP4 holding only the range [start, end) of one clip. Timeline compose takes one still and one video and returns one MP4 with both on screen. Video captions accept authored cues with text, start and end, so the on-screen words are yours. These are mechanisms; whether a given edit is creative enough is the Office's case-by-case question, not something a route decides.

Do these steps make a video protected?

The report does not say so, and neither do the Sume docs. What the report does say is that the arrangement and modification you author are the kind of contribution that can count. Keeping your own drafts, your selection of takes, edit decisions and the text you wrote gives a record of what you did.

How would I keep an edit record?

Save the request bodies you send for each edit, with the start and end values and the cue text, next to the job ids. For longer pieces, see assembling long-form video with Timeline and trim, filter and detach audio. Ask counsel about any registration or licensing decision.