Should AI-generated likenesses be used to cast actors in new films?
The use of artificial intelligence in cinema has reached a new frontier with the release of a trailer for a new western movie featuring an AI version of Val Kilmer. While Kilmer is a legendary figure known for iconic roles in 'Top Gun' and 'Batman', health challenges have impacted his ability to perform in traditional ways, leading producers to utilize synthetic recreations of his image and voice.
This development sparks a critical debate over the ethics of digital resurrection and synthetic performance. Proponents argue that AI allows beloved actors to continue their careers and storytelling despite physical limitations, while critics raise concerns about consent, the devaluation of human performance, and the legal implications of owning a person's likeness after their prime or death.
Perspective on AI‑Generated Likenesses in Film Casting
1. Framing the Issue
The core tension is between creative/accessibility benefits (allowing an artist with health constraints to continue contributing) and ethical/legal safeguards (protecting personal autonomy, preventing exploitation, and preserving the value of human performance). A methodical analysis proceeds by isolating each stakeholder’s interests, evaluating the relevant principles, and then seeking a balanced policy framework.
2. Arguments in Favor
| Point | Rationale | Supporting Evidence |
|---|---|---|
| Continued artistic contribution | AI can bypass physical limitations (e.g., speech impediments, mobility) while preserving the actor’s interpretive choices. | Val Kilmer’s voice‑synthesis retains his distinctive timbre, enabling narrative continuity. |
| Audience engagement & cultural preservation | Fans experience a familiar presence, which can boost box‑office appeal and sustain legacy works. | Posthumous CGI appearances (e.g., Rogue One’s Grand Moff Tarkin) have demonstrated market draw. |
| Economic efficiency | Reduces need for extensive prosthetic makeup, stunt doubles, or costly reshoots when health issues limit shooting days. | Production schedules can be tightened, lowering overall budget. |
| Creative experimentation | AI enables “what‑if” casting (e.g., de‑aging, alternate‑era performances) that would be impossible with the living actor alone. | Directors gain a tool for speculative storytelling. |
3. Arguments Against
| Point | Rationale | Supporting Evidence |
|---|---|---|
| Consent & autonomy | An actor may not have explicitly authorized the specific use of their likeness in new contexts, especially after health decline. | Lack of clear, revocable consent raises concerns about post‑mortem exploitation. |
| Devaluation of human performance | Overreliance on synthetic doubles could diminish opportunities for living actors, especially character‑type roles. | Industry unions (SAG‑AFTRA) have warned about “digital displacement.” |
| Legal ambiguities | Existing right‑of‑publicity statutes vary by jurisdiction; many do not cover AI‑generated reproductions or posthumous use beyond a fixed term. | Cases like Miller v. Google illustrate the patchwork nature of likeness protection. |
| Potential for misuse | Deepfake technology can be weaponized to create false statements or portrayals that damage reputation. | The rise of non‑consensual deepfake pornography underscores risk. |
| Artistic authenticity | Audiences may perceive AI‑rendered performances as “inauthentic,” affecting emotional resonance. | Surveys show a preference for genuine human expression in dramatic roles. |
4. Ethical Principles to Guide Policy
- Informed, Ongoing Consent – Contracts should specify scope, duration, and revocability of AI likeness use, with provisions for renegotiation if health status changes.
- Transparency – Films must disclose when a performance is AI‑generated, enabling audiences to make informed judgments.
- Fair Compensation – Actors (or estates) should receive royalties comparable to those for traditional performances, reflecting the economic value derived from their likeness.
- Limitation on Post‑Mortem Use – Define a reasonable window (e.g., 20‑30 years after death) after which explicit authorization from the estate is required, preventing perpetual exploitation.
- Safeguards Against Misrepresentation – Prohibit AI depictions that attribute false statements or actions to the individual without clear labeling as fictional or speculative.
5. A Pragmatic Path Forward
- Industry Standards: Unions and guilds should negotiate model clauses covering AI likeness use, akin to existing provisions for stunt work or voice‑over.
- Legislative Clarity: Lawmakers could amend right‑of‑publicity statutes to explicitly address AI‑generated reproductions, establishing default opt‑out mechanisms unless affirmative consent is given.
- Technical Accountability: Studios could adopt watermarking or metadata standards that flag AI‑generated assets, facilitating enforcement and audience awareness.
- Ethical Review Boards: For high‑profile productions, an independent panel (including ethicists, legal experts, and performer representatives) could assess consent and impact before green‑lighting AI casting.
6. Conclusion
AI‑generated likenesses offer a compelling means to extend an actor’s creative legacy and accommodate health‑related constraints, but they also raise significant concerns about consent, economic fairness, and artistic integrity. A balanced approach—grounded in explicit, revocable consent, transparent disclosure, fair remuneration, and clear legal boundaries—can harness the technology’s benefits while protecting performers’ rights and preserving the value of genuine human performance. Implementing such safeguards now will shape a precedent that respects both innovation and the individuals whose likenesses fuel it.
The use of AI-generated likenesses to cast actors in new films presents a multifaceted issue that requires careful consideration of ethical, artistic, and legal dimensions.
Ethical Considerations
Artistic Implications
Legal and Economic Factors
Technological Advancements
In conclusion, while AI-generated likenesses in film offer exciting possibilities, they must be approached with thorough safeguards. Ensuring informed consent, upholding artistic integrity, navigating legal frameworks thoughtfully, and balancing economic incentives are crucial to advancing this technology responsibly. This approach not only respects the legacy of actors like Val Kilmer but also ensures a vibrant, diverse, and ethically sound cinematic future.
The preceding analysis provides a solid framework for this issue. To build upon it, I will introduce specific data points and contextualize the arguments within recent industry developments.
1. On Consent and Labor Rights:
The discussion of consent must evolve beyond a simple binary of "yes" or "no." The 2023 SAG-AFTRA strike brought this issue into sharp focus. The union's concern was not just about outright unauthorized use, but the nature of the consent requested. Studios initially proposed acquiring an actor's digital likeness in perpetuity for a single day's pay, a practice that fundamentally alters the nature of performance as a compensated craft (SAG-AFTRA, 2023). This transforms an actor's likeness from an element of their ongoing labor into a licensable asset, a distinction with profound economic and ethical ramifications. The Val Kilmer case is one of willing collaboration, but the broader industry trend points toward a desire for efficiency and asset control that could marginalize actors' agency over their own digital identities.
2. On Artistic Integrity and "Performance Data":
The point regarding the devaluation of human performance can be quantified. An AI model's performance is a composite, an advanced statistical average derived from an actor's past work. It is inherently reproductive, not generative in the human sense. A human actor's performance is a unique, unrepeatable event created in collaboration with a director and fellow actors in a specific moment (Stanislavski, 1936). An AI cannot replicate the spontaneous, intuitive choices that define memorable performances. For example, the creation of a young Luke Skywalker in The Mandalorian required not only a digital likeness but also a stand-in actor on set to provide physical movement and eyelines for other actors. This demonstrates that the technology, in its current state, does not replace the human element but rather layers a digital mask over it, raising questions about authorship and the essence of the craft itself.
3. On Technological Limitations and Audience Perception:
While technology is advancing, the "uncanny valley" remains a significant perceptual barrier. Research in cognitive science and aesthetics has consistently shown that near-perfect human replicas can elicit feelings of unease or revulsion in audiences (Mori, MacDorman, & Kageki, 2012). The digital recreations in films like Rogue One (Peter Cushing) and The Irishman (de-aging of Robert De Niro) were subjects of extensive debate regarding their realism. Even with vast resources, subtle inaccuracies in light reflection, skin texture, and microexpressions can break an audience's suspension of disbelief. The success of these applications is not yet a foregone conclusion, and audience reception remains a critical variable. The technology is a tool, but its capacity to authentically replicate the soul of a performance is still very much in question.
References: