Edition 006 — The Person Inside the Asset

The easiest way to misunderstand creator content is to treat it as a flat image.

It is often closer to a layered file.

There may be a script, photograph, video, music track, graphic treatment, caption, logo, product, voice, face, performance, and edit history. Different people may have contributed to those layers. Different elements may have different authors, owners, licenses, consents, or approval terms. A brand may be authorized to publish the approved whole without being authorized to change every part, repurpose the person’s identity, or claim authorship of every contribution.

This is why four familiar sentences fail so often:

“You said we could post it.”
“We credited you.”
“We paid for it.”
“It was work made for hire.”

Each may describe something real. None should be asked to do the work of the others.

Permission to post is not a blank editing pass

Imagine a creator approves a warm, quiet product video. The brand later crops it, accelerates it, replaces the music, changes the caption, adds a price graphic, and places a call to action over the creator’s face.

The underlying file may be the same. The communication is not.

Subject to the limitations elsewhere in federal copyright law, the copyright owner has the exclusive right to prepare derivative works based on the copyrighted work. A derivative work is one that recasts, transforms, or adapts a preexisting work. But naming that right does not turn every crop, resize, caption, translation, or technical adjustment into an automatic legal conclusion. Whether a particular change implicates the derivative-work right depends on what was changed, how it was used, what rights were granted, and what other rules apply.

The production question is still practical: which changes were already authorized, which require approval, and which are outside scope?

An agreement can identify different lanes:

technical changes needed for delivery;
agreed platform crops, resizing, and captioning;
approved cutdowns, accessibility treatments, or translations;
substantive alterations requiring creator review; and
prohibited changes that would create a new endorsement, meaning, or context.

The goal is not to make every resize a negotiation. It is to prevent a necessary resize and a newly manufactured message from being treated as the same act.

Dove’s campaign archive describes Evolution as revealing the construction of a beauty image. That makes it a useful cultural prompt—not evidence about an unrelated creator’s contract. Selection, retouching, sequencing, and context can alter what an image communicates. The lesson is not that every edit is forbidden. It is that an approved file and an approved message are not necessarily the same thing.

A useful workflow makes the difference visible before the edit reaches an audience. The editor knows which changes can be made without another round of review. The creator knows which changes will return for approval. The brand knows which placements or messages would require a new conversation. That is not friction added to the work. It is production information the work already needed.

It also gives approval an object. Instead of “the creator approved the campaign,” the record can identify the exact file, version, caption, placement, territory, and date that were reviewed. When the next edit arrives, the team can compare it with something concrete. The question becomes whether the new version remains inside the recorded lane—not whether somebody can reconstruct an old conversation from a message thread after the campaign has already changed.

Permission and platform enforcement are also different layers. A valid license may not prevent an automated claim, so preserve the exact license, asset identifier, approved use, and version record needed to explain the use when a platform asks.

Credit is a label, not a permissions system

“Credit to the owner” is one of the internet’s most persistent magical phrases.

Credit can be ethically important, contractually required, professionally useful, or legally relevant in particular contexts. It can help audiences find the creator and preserve the history of a work. But naming a source does not, by itself, establish permission, ownership, payment terms, or a copyright exception.

The dispute involving an Associated Press photograph made by Mannie Garcia and Shepard Fairey’s HOPE image is useful only with care. Fairey used the AP photograph as source material. The later civil copyright case involved contested positions and eventually settled. Because it settled, it did not produce a final civil infringement judgment that resolves the broader permission question. Fairey’s separate criminal case concerning evidence misconduct should not be confused with a civil merits ruling about the image.

That history matters precisely because it refuses the easy version of the story. Recognizing a source, identifying its maker, and understanding the legal basis for a later use are related tasks. They are not interchangeable ones.

Before reposting or adapting material, separate four boxes:

Source: Who created or published this exact material?
Permission or exception: What authorizes this use?
Credit: What attribution is required or appropriate?
Commercial terms: Is payment, licensing, or another condition involved?

Filling one box does not silently fill the other three.

This distinction becomes especially important in fast-moving campaign work. A social team may identify the correct source but not know whether a platform repost, paid advertisement, email placement, or edited cutdown is authorized. A creator may be fully credited but still disagree with the new context. A brand may have paid for one deliverable without settling every later use. The boxes force the team to name the missing question instead of hiding it inside the word “credit.”

The person is not just another layer of content

An on-camera creator contributes more than a video file. Their face, voice, name, performance, reputation, and relationship to the audience may carry commercial and personal significance.

Depending on the facts and jurisdiction, a campaign may raise overlapping questions involving copyright in the script or recording; contractual permission to use the approved content; use of a name, image, likeness, or voice; endorsement and advertising; privacy; publicity law; or union obligations. These bodies of law are not interchangeable. Identity-related rights and remedies are jurisdiction-specific, so a useful public framework should teach issue-spotting rather than pretend to supply one national answer.

The 2024 public dispute between Scarlett Johansson and OpenAI over the “Sky” voice illustrates the distinction without resolving it. OpenAI said Sky was voiced by a different professional actor, was not intended to resemble Johansson, and had been selected before the company contacted her. In a public statement carried by the Associated Press, Johansson said she had declined an offer and disputed the circumstances surrounding the voice.

Those are competing attributed accounts, not findings from a judgment cited here.

The episode still exposes a practical question. Authorization to use one performer’s recording does not necessarily answer every question about perceived identity, association, or endorsement. A company may believe it cleared the recording in front of it while an audience hears somebody else in the result. That does not decide liability. It tells the production team that “Who performed this file?” and “Whose identity might the audience perceive?” are not always the same question.

Before a campaign expands, ask:

Which approved content may be used?
Is the creator’s name or handle part of the use?
May the creator’s face or voice be extracted, simulated, edited, translated, or paired with new copy?
Would the new placement imply a different endorsement or association?
Who approves materially changed versions?

In an audience’s experience, the content and the person may be difficult to separate. The production record should not assume they are legally or operationally identical.

That record does not need to be theatrical. It can be a short approval table attached to the asset: approved file, approved placements, permitted edits, identity uses, prohibited contexts, approver, and version date. The point is to preserve the person inside the asset when the asset begins moving through a campaign system.

“Paid work” is not the work-made-for-hire test

Few copyright phrases are used more casually than work made for hire.

Under federal copyright law, there are two routes. The first is a work prepared by an employee within the scope of employment. The second is a specially ordered or commissioned work that fits one of nine uses listed in the statute and is covered by an express signed written agreement stating that it is a work made for hire.

Payment alone does not satisfy that definition.

For the employee route, courts apply common-law agency principles to the actual relationship. In Community for Creative Non-Violence v. Reid, the Supreme Court identified factors including control over the work, the skill required, the source of tools, the location and duration of the relationship, assignment of additional projects, discretion over working time, method of payment, the role in hiring assistants, employee benefits, and tax treatment. No single label or factor replaces the analysis.

For the commissioned-work route, the intended use matters. A photograph, illustration, video, or social post is not automatically included or excluded because of its medium or because an invoice was paid. The work must be specially ordered or commissioned for use within one of the statute’s listed categories, and the parties must expressly agree in a signed writing that it is a work made for hire.

The consequence is substantial. For a qualifying work made for hire, the employer or other person for whom the work was prepared is considered the author for purposes of federal copyright law. Section 201(b) separately provides that person owns the copyright unless the parties expressly agree otherwise in a signed written instrument.

Classification also changes the clock. For covered works created on or after January 1, 1978, a work made for hire generally receives a term of 95 years from first publication or 120 years from creation, whichever expires first. Section 203’s mechanism for terminating certain post-1977 grants applies to grants involving works other than works made for hire.

This does not mean clients cannot obtain broad or complete rights in commissioned work that does not qualify. Assignment and licensing are separate tools. It means the parties should use the right tool and describe it accurately.

The Copyright Office’s Circular 30 turns the definition into a useful decision path: employee and scope of employment; or an express signed agreement plus an eligible commissioned use. That structure is far more useful than the sentence “we paid, so we own it.”

Read the asset in layers

Before approving a use or edit, run a non-exhaustive production check:

Work: What protected expression is present, and who contributed it?
File: What exact deliverable or source material was supplied?
Source package: Were raw, layered, project, or other editable files included—or only the approved final export?
Permission: Which acts and changes were authorized?
Credit: What attribution is required or appropriate?
Person: Whose identity, voice, likeness, performance, or endorsement appears?
Relationship: Employee, independent contractor, collaborator, platform participant, or something else?
Approval: Who must review materially changed versions?

A 15-second asset can contain several separate questions. Making the layers visible early gives the parties a chance to design a sane workflow before the work, the file, and the person are pulled apart by production.

Sources and further reading

General educational information, not legal advice. Copyright, contract, privacy, publicity, advertising, labor, and platform issues depend on the facts and jurisdiction.