The Ethics of Synthetic Talent: Navigating the New Frontier

The first wave of digital human advertising arrived mostly as spectacle. Brands showcased the technology because it was new, because it generated press, because audiences were startled by how lifelike synthetic performers had become. The question asked was "can we?" — and the answer was increasingly yes. The question rarely asked was "should we, and under what conditions?"

That second question is now unavoidable. Regulators are paying attention. Audiences are developing nuanced opinions. Platform policies are evolving. And the brands that navigate this landscape thoughtlessly are discovering that the reputational cost of getting it wrong can far exceed any production savings the technology delivers.

At Variete Productions, this is a conversation we have with every client who asks about synthetic talent. Here is the framework we use.

First: Distinguish the Different Use Cases

Synthetic talent is not a single thing. The ethical weight attached to each use case is quite different, and conflating them produces muddled thinking:

• Fully synthetic personas. A digital character created from scratch, with no real human as its basis. Think of an AI-generated brand mascot or virtual influencer who has never existed as a real person. No likeness is being used, no consent is required, no real human's identity is in play. This is the most ethically uncomplicated application.
• Digitally extended real talent. A real actor or model who has been filmed, and whose performance is then extended, de-aged, translated into a different language, or placed in environments that were never shot. This requires explicit contractual consent from the talent, clear scope definitions, and fair compensation. Without those, it is exploitation.
• Likeness replication without consent. Using AI to generate the appearance of a real person — a celebrity, a public figure, even a private individual — without their knowledge or agreement. This is not a grey area. It is legally actionable in most jurisdictions and ethically indefensible regardless of legality.
• Historical or deceased figures. Using AI to resurrect the likeness of people who cannot consent — historical figures for educational content, deceased performers for posthumous campaigns. This requires exceptional care, strong ethical justification, and in commercial contexts, the explicit authorization of the estate.

The Transparency Standard

Our position at Variete Productions is that audiences have a right to know when they are watching synthetic content, particularly when that content features human-like figures in commercial contexts. This is not a legal requirement in all jurisdictions yet — but it is becoming one rapidly, and more importantly, it reflects a basic standard of honesty that we believe brands should want to uphold.

Practically, this means: campaigns using synthetic talent should include disclosure — at minimum in accompanying platform metadata, and ideally in the creative itself where appropriate. Some clients worry that disclosure undermines the impact of the content. Our experience is the opposite. Audiences who know they are watching a genuinely impressive technical achievement respond with curiosity and engagement. Audiences who later discover they were misled respond with distrust that is very difficult to recover from.

"The brands that will own this space long-term are the ones that use AI boldly and transparently — not the ones that use it to deceive. Audiences are smarter than the industry gives them credit for." — Michal Jaworski

Protecting Real Talent in the AI Era

For productions involving real performers whose work will be extended or modified by AI, we use what we call a Synthetic Performance Agreement — a contract addendum that specifies in plain language exactly what AI processing will be applied to the footage, what the outputs will be used for, how long the rights extend, and what compensation structure applies to AI-derived uses. This protects the talent. It also protects the brand, because productions built on clear agreements are not at legal risk when the regulatory landscape tightens — which it will.

The Screen Actors Guild agreements of the past several years have begun to formalize some of these protections at the industry level. Brands working with non-union talent should not take that as license to disregard the same principles. The protections exist because they reflect real ethical stakes, not just because a union negotiated for them.

Synthetic Environments vs. Synthetic People

It is worth separating two things that often get grouped together in discussions of AI video. Synthetic environments — AI-generated backgrounds, set extensions, atmospheric effects, abstract visual worlds — carry almost none of the ethical complexity attached to synthetic people. No one's likeness is used. No consent is required. The creative is clearly a construction. This is the majority of what we produce at Variete, and it is essentially uncontroversial from an ethics standpoint.

The difficult territory is specifically human likeness. That is where brands need clear thinking, clear contracts, and clear communication — both internally and with the audiences they serve.

A Framework for Decision-Making

When a client comes to us with a synthetic talent brief, we run through four questions before production begins:

• Is any real person's likeness being used or approximated? If yes, do we have documented consent?
• Is the use consistent with how that person has positioned themselves publicly and professionally?
• Would the audience, if fully informed, feel they had been deceived — or impressed?
• Does this creative serve the brand's long-term reputation as well as its short-term campaign goals?

If the answers are satisfactory, we proceed with enthusiasm. The technology is remarkable, the creative possibilities are genuinely exciting, and when used with integrity, synthetic talent can produce work that would be impossible any other way. The goal is to use it in a way that you would be comfortable defending publicly — because eventually, you will be asked to.