Most studios, agencies and media companies now have AI somewhere within their pipelines and workflows. That is no secret. What the creative industry still resists is AI disclosure, saying openly where and how that use shows up in the work.
The reasons are understandable. Agencies fear that disclosure will give procurement teams a reason to reduce fees. Studios expect legal teams to slow delivery because of rights and liability concerns. Leadership teams worry that employees will interpret AI adoption as a redundancy signal, while brands anticipate audience hostility towards anything labelled AI-generated.
Each concern is a real commercial risk. What has changed is the ground they’re standing on. When 58% of creative professionals say they have used AI without telling a client, non-disclosure is no longer an occasional exception (Envato State of AI in Creative Work 2026, n=1,780). It has become part of normal industry practice. The question has shifted from whether AI is being used to whether the business can explain where it was used, what it contributed, and who remains accountable for the finished work.
When AI reaches the final frame
For a time, non-disclosure carried little obvious cost. When AI helped with research, previs storyboards, internal references or early-stage idea development, it remained inside the production process. The client was still buying the final creative judgement, the team’s expertise and the finished asset.
The position changes when AI affects what the client or audience sees or hears in the final output. A synthetic performer in an advert, generated footage in a case film, a replicated voice, a trained likeness or an entirely generated scene all require a clear production record.
At that point, a business that cannot explain how the work was made is no longer occupying a neutral position. It is exposed. In the conversations we’re having with leaders across the industry, the same reservations repeatedly arise. Each is legitimate, but none is resolved by silence.
The fee problem is pricing, not disclosure
Only 27% of agencies believe they are paid fairly for their work, and the industry’s heavy reliance on time-based pricing is the reason (IPA). That pricing model is the real exposure when AI enters the conversation, not the disclosure. Agencies fear that revealing AI use gives procurement a reason to ask why the work still costs the same when it can be done faster.
That concern is valid when production is still priced primarily according to time. Under that model, a faster method can appear to justify a lower fee. The underlying issue is not disclosure. It is a pricing model that fails to capture the value of judgement, craft and results.
AI brings that existing tension to the surface, whether businesses discuss it openly or not.
If a task takes four hours instead of four days, the agency still needs to price the creative direction, the rights position and responsibility for the outcome. Being open about the method allows that commercial conversation to happen properly, rather than postponing it until the client discovers the process elsewhere.
The production record is the IP defence
Intellectual property is next. AI creates genuine legal uncertainty, so keeping its use private can feel safer. That position holds only until a client, rights holder, or regulator asks how the work was produced. A business that can point to a clear production record is in a far stronger position than one trying to reconstruct events after delivery. That record should show what was generated, filmed, licensed, prompted or manually rebuilt, and what entered the finished asset. Openness does not create IP exposure. The absence of a defensible record does.
This is also why AI governance cannot sit solely with legal or compliance teams. Someone within the production process needs to understand the creative work, the technology and the rights implications well enough to document decisions as they happen. The same provenance record is turning up in hiring, where senior creatives now submit evidence chains rather than polished portfolios.
Consent is where the exposure becomes real
Talent consent carries even greater weight. The assumption is often that raising AI usage with performers, artists or contributors will complicate production. In practice, late consent creates greater delay and risk. Performer likeness, synthetic casting, voice replication and training rights need to be included in the production plan from the outset. They cannot be resolved retrospectively after a model has been trained or an asset has been delivered.
This is not solely a European issue. New York now requires certain adverts to identify the use of AI-generated synthetic performers. California is introducing detection and provenance obligations for synthetic images, video, and audio, while the FTC is treating AI-enabled deception as a consumer protection issue.
The rules differ across the US and Europe, but the direction is consistent. Agencies and studios increasingly need a written record showing what was generated, what was performed, what was licensed and who approved it. That record only exists when the conversation begins early enough.
Silence forces employees to guess
Then there is internal anxiety. Leadership teams often avoid discussing AI because they don’t want employees to assume automation will lead directly to job losses. The difficulty is that silence rarely reduces that anxiety. It usually intensifies it.
Employees can already see tools entering production, tasks changing, and job descriptions asking for AI literacy, ComfyUI, and provenance awareness. When leadership says nothing, people fill the gap themselves, and their conclusions are often more alarming than the reality.
A credible internal message should explain what is changing, which decisions still require human judgement, which roles need new capabilities, which risks the company will not take and where accountability will remain. That will not eliminate every concern, but it gives employees something more useful than rumour.
Name the human, and the disclosure becomes credit
Public reaction is the concern most often cited. Some audiences will object to AI-created work, and a blunt or overly broad disclosure creates more confusion than clarity. The scale of AI involvement varies enormously. A minor retouching assist sits at one end, a fully synthetic performer at the other. Treating the two as equivalent turns disclosure into an unhelpful warning label. The language needs to describe what actually happened. It should also identify the human contribution.
Who developed the idea? Who made the creative judgements and stood behind the final image? When the human role disappears, disclosure sounds like an admission. When it is named clearly, disclosure becomes part of the production record and the credit.
Awards and regulators have already stopped waiting
DM9 lost a Cannes Lions Grand Prix in 2026 after an undisclosed, AI-manipulated case film simulated real-world campaign outcomes, and Cannes now mandates disclosure of AI involvement at entry (Adweek, 2026). The awards industry was among the first parts of the creative sector to act. Nobody lost a Grand Prix for using AI. They lost it for not saying so.
D&AD has also incorporated disclosure into its 2026 entry process through an Entry Validation Card signed by a senior representative, alongside declarations covering AI involvement across categories.
This matters beyond the awards themselves. Awards influence pitches, recruitment, agency reputations and senior creative careers. Once AI disclosure becomes part of the entry process, it also becomes part of the commercial record. The case film, production trail and credits must all align. Regulation is moving in the same direction. The EU AI Act provides one of the clearest deadlines. From 2 August 2026, people in the EU must be informed when they interact with certain AI systems or are exposed to particular forms of AI-generated or manipulated content.
The practical challenge for creative and media businesses is not simply placing a label at the end of the work. It is knowing which label is accurate. That requires AI to be discussed much earlier, within the brief, production plan, talent agreements, client approvals and archive. When left until final delivery, disclosure becomes a hurried attempt to reconstruct the process. When built into production, it becomes routine.
AI-proficient talent leaves for open competitors
There is a real cost to hiring to keep AI work off the record. The people who can make disclosure workable, senior creative technologists and producers who understand the process, the rights and the client conversation, are exactly who a business needs. A business that treats its work as something to keep hidden will struggle to attract them and will tend to lose the ones it has to competitors already operating openly.
Silence is a communication choice
Much of this comes down to communication. What should a business say publicly about its use of AI, and who has the authority to decide? For most, the current answer is nothing and no one. That is still a position, just an unmanaged one.
A business that cannot explain its use of AI to a client, jury, employee, or audience has a messaging problem before it has a legal or talent one. The businesses moving ahead have determined what they will disclose, what they will not, how they will describe different levels of AI involvement and how they will make the human role visible. They have also decided who is responsible for approving those statements.
For smaller teams without a legal function, that might mean a fractional advisor or a production lead holding the brief rather than a dedicated hire from day one. The capability gap is real, but it is a gap to close rather than a reason to stay quiet.
Trust now sits in plain answers
The strongest creative and media businesses will not be those that treat AI as something to keep quiet. They will be those capable of separating tool use from authorship, and production speed from commercial value.
AI can sit almost anywhere in the pipeline, as long as the business can say where it sits, what it changed, and who remains accountable. That is where trust now sits.
Which turns the whole thing into a hiring question. The person capable of leading that conversation, who understands the work, the technology, and what a client, jury, employee, or regulator is likely to ask, either already exists within the business or is its next important senior appointment. We made the same case about what’s really preventing AI adoption across creative and media, leadership and governance, not the tools.
