What Ownership Would Actually Mean
Once current law is seen as a patchwork, the next question is what a federal digital-replica right would actually control. This chapter examines the Copyright Office’s 2024 recommendation: how narrowly to define a protected replica, whom to protect, when liability should begin, and how a person’s consent should survive a contract, a change in circumstances, or death. The useful judgment is not whether the proposal sounds strong, but whether each design choice gives people control without turning every resemblance into a veto over expression.
The Office’s report is a recommendation, not binding law. Its value is that it makes the hidden choices visible. A federal statute would have to decide what counts as the relevant object, what conduct causes legally significant harm, what a distributor must know, and whether permission is a permanent transfer or a bounded relationship.
A Narrow Right Aimed at a Public Act
The Office recommends a subject narrower than a general right in a person’s name, image, or likeness. It would target replicas so realistic that they are difficult to distinguish from an authentic depiction. That threshold responds to the distinctive harm of a synthetic recording that appears to show the real person speaking, singing, appearing in an advertisement, or participating in an event. It would not automatically cover every caricature, catchphrase, visual resemblance, or artistic evocation that reminds an audience of someone.
The difference is practical. A comedian’s exaggerated imitation may refer to a celebrity without pretending to be an authentic recording. A historical drama may use digital techniques to depict a person in a scene that never occurred, while signaling that the work is a reconstruction or fictionalization. By contrast, a fabricated clip designed to look like an authentic statement asks the audience to treat an invented act as evidence about a real person. The proposed threshold tries to protect against that identity-level deception without giving one person ownership over every cultural reference to them.
The Office also recommends protecting all individuals, not only people who are famous or whose identities have commercial value. That choice follows from the kinds of harm the report identifies. A cloned family voice used in a ransom scam does not depend on celebrity recognition. A private person can be placed in an explicit image, bullied with a fabricated video, or falsely presented as participating in conduct that damages their relationships. Fame may increase the audience for a replica, but it is not what makes the underlying injury serious.
Universal coverage also avoids a difficult threshold question: how famous must someone be before their face or voice becomes valuable enough to protect? If protection depends on market value, a person may need to prove the very commercial status the law is supposed to make unnecessary. The Office’s recommendation instead treats the identity interest as personal and widely shared, while keeping the subject matter narrow through the realism requirement.
The next design choice is the point in the lifecycle where a legal rule should attach. The Office recommends covering publication, distribution, public performance, public display, or making an unauthorized replica available to the public. These are the acts most likely to create continuing harm. Once a synthetic image or voice is released, others can see it, copy it, and rely on it. The public release is therefore more important than the mere existence of a file on a private device.
The recommendation does not treat creation alone as automatically actionable. Private experimentation or personal entertainment may be innocuous and can support creative work. Someone might test a voice model privately to understand what the technology can do, or create a draft that is never shown to anyone. Making that act liable by itself could turn the right into a restriction on experimentation rather than a response to public harm.
That boundary does not make creation irrelevant. If a person knowingly creates a replica as part of a distribution scheme, the creation can be evidence of the larger conduct. Other criminal or civil laws may also apply to especially harmful material. The point is to distinguish a private act that remains contained from a public act that makes the replica part of the information environment.
The Office recommends an actual-knowledge standard for direct liability. A distributor would need to know both that the representation was a replica of a particular individual and that it was unauthorized. This is more demanding than a rule asking what the person “should have known.” The distinction matters because a user may not recognize that an image or recording is synthetic, may not identify the person being imitated, or may reasonably believe that permission was obtained.
At the other extreme, the Office does not require an intent to deceive. That requirement would miss uses that are not designed to fool the audience but still exploit or injure the person depicted. A replica used to harass, ridicule, or profit from a performer’s voice may be unauthorized even if the audience understands that it is artificial. Requiring proof of a subjective plan to deceive would make protection depend on an inner motive that is difficult to establish and narrower than the harm.
The result is a three-part boundary. The protected subject is a realistic identity replica, not every resemblance. The protected conduct is public dissemination, not every private act of creation. The knowledge requirement asks what the distributor actually knew, not merely what they might have discovered and not only whether they meant to deceive. Each boundary narrows the right in a different direction: the first limits the object, the second limits the act, and the third limits the mental state.
Consent Has a Lifecycle
A right that follows a person’s face or voice cannot be useful if authorization is either impossible or permanent by default. The Office recommends that individuals be able to license their digital-replica rights, allowing them to participate in creative products, services, and compensated uses. But it recommends against allowing a person to assign all of those rights outright. The proposed structure treats identity as a hybrid: it can carry economic value like property, but it also involves privacy and personal control that should not be sold away forever.
The distinction between a license and an assignment is easier to see in practice. A license gives another party permission to use a voice or likeness under agreed terms. It can identify the project, medium, duration, territory, purpose, compensation, and limits on new uses. An assignment transfers the rights themselves, leaving the recipient to control them going forward. The Office’s concern is that a person could sign a broad assignment early in a career, under unequal bargaining power or without understanding how synthetic media might evolve, and later lose the ability to object to uses that were never meaningfully contemplated.
The report suggests that non-collectively-bargained licenses might be limited to a relatively short period, such as five or ten years. That period is an example for legislative design, not a fixed rule already in force. The underlying mechanism is renewal: when circumstances change, the parties must return to the bargaining table instead of allowing an old agreement to function like perpetual ownership. A performer may learn that a new use damages their reputation; a product may change markets; or a technology may make a previously narrow permission much more powerful than expected.
A time limit also creates a transition problem. Suppose a singer authorizes a label to use a digital replica of their voice in recordings for several years. If the license ends, it would be disruptive and potentially unfair to require every lawfully produced recording or package to disappear immediately. The Office therefore distinguishes new uses from continued distribution of works prepared during the lawful term. The license would block new replica uses after expiration, while material created under the agreement could continue to be used according to the terms that governed it.
The lifecycle can be represented as a series of transitions rather than a single ownership event:
The diagram does not describe current enacted law. It shows the control problem the Office wants legislation to solve: permission has a scope and a term; an existing work may outlast the permission to create new material; a minor may need extra protection; and postmortem control is a separate policy question rather than an automatic consequence of calling the right “property.”
Informed consent is what makes the transition into the licensed state meaningful. The Office points to collective bargaining provisions that require digital-replica terms to be clear and conspicuous, set apart in a rider or another form that stands out. A person may technically sign a contract while not understanding that a clause authorizes a synthetic voice to generate new lines, new performances, or advertising in unrelated contexts. A prominent, specific agreement makes the permission easier to evaluate before the replica is made and easier to interpret after a dispute begins.
This concern is especially strong for minors. Children and young performers generally have less bargaining power and less ability to understand a long-term identity grant. The Office considers safeguards such as court review, protected earnings, and licenses that expire when the minor reaches adulthood. These mechanisms change the timing of control. Instead of treating a child’s signature or a guardian’s agreement as an irrevocable decision about an adult identity, they create a point at which the person can reassess the permission with greater capacity and information.
Postmortem rights expose a different tension. The Office’s primary recommendation is that a federal right should prioritize working artists, the dignity of living people, and public protection from fraud and misinformation; for those purposes, it says a postmortem term is not necessary. At the same time, the report recognizes a reasonable argument for letting heirs control and benefit from a deceased person’s commercially valuable persona. If Congress creates such a right, the Office suggests an initial term shorter than twenty years, with a possible extension while the persona continues to be commercially exploited.
That disagreement is not a side issue. A long postmortem term can protect an estate’s investment and prevent unauthorized uses that distort a performer’s legacy. It can also give heirs or corporate successors control over expressive depictions of someone who can no longer consent or object, and it can make it difficult to identify the person who is entitled to authorize a use. The right’s duration therefore changes not only who receives income, but also who can decide how a person appears in public after death.
The Office’s proposed lifecycle is an attempt to keep those interests from collapsing into one label. The person remains the starting rights-holder. A licensee receives permission, not necessarily permanent ownership. A contract should make the permission legible. Time can require renegotiation. A minor may need a later checkpoint. An existing work may persist after a license expires, while new uses require a fresh agreement. Postmortem control, if provided, may be shorter and more conditional than a general property term.
These safeguards also reveal the proposal’s practical theory of ownership. Ownership is not only the power to say yes or no to a file. It is the power to set the circumstances under which a recognizable identity may be made to speak, sing, appear, or endorse. A right that cannot be licensed would prevent beneficial uses and compensation. A right that can be assigned without limits could leave the person with no meaningful future control. The Office’s middle position tries to preserve both agency and lawful creative use.
The unresolved question is enforcement at the moment a replica becomes public. A rule aimed at distribution and actual knowledge may avoid sweeping in private experimentation, but it must still work when the uploader is anonymous, the platform is hosting the material, or the replica appears inside a work that claims expressive protection. The next stage of the problem is therefore operational: how can a person obtain removal and a remedy without making intermediaries responsible for deciding every disputed act of speech?