For the past several years, most conversations about generative AI have focused on what the technology can create.

The next phase will focus just as much on what happens after that content is generated:

Which model created it? What source assets were used? Was the content substantially manipulated? Does it depict a real person? Can its production history be verified? And who is responsible for disclosing its artificial origin when it is published?

These questions are becoming increasingly important for studios, agencies, brands, broadcasters, and other professional creative teams. They are also becoming part of the regulatory environment.

On August 2, 2026, Article 50 of the European Union’s AI Act became applicable. It introduces transparency obligations for providers and professional deployers of certain AI systems, including systems that generate or manipulate images, video, audio, and text.

At Sequencer, we do not see this simply as a compliance exercise. We see provenance and governance as necessary infrastructure for professional AI production.

The EU AI Act does not prohibit studios or companies from using AI to generate images and videos.

Instead, it creates clearer responsibilities for the different participants involved in creating and publishing AI-generated content.

Under Article 50(2), providers of AI systems that generate synthetic image, video, audio, or text content must ensure that relevant outputs are marked in a machine-readable format and are detectable as artificially generated or manipulated. Those technical measures must be effective, interoperable, robust, and reliable, as far as technically feasible.

This technical requirement is distinct from placing a visible “AI-generated” notice over every image or video.

Machine-readable marking is intended to make the origin of content detectable by platforms, customers, verification systems, and other authorized participants. Depending on the implementation, this may involve provenance metadata, content credentials, digital signatures, cryptographic references, or other technical signals connected to the asset.

The Act also recognizes that not every use of AI should be treated in the same way. Article 50(2) provides an exception when an AI system performs an assistive function for standard editing or does not substantially alter the customer’s original material or its meaning.

That distinction matters in professional production.

Resizing a video, adjusting compression, reducing noise, or performing basic color correction is fundamentally different from generating an entire scene, replacing a person’s face, cloning a voice, extending an environment, or turning a still image into a synthetic video.

A professional production platform therefore needs to understand not only whether AI was involved, but how it was involved.

One of the most important distinctions in the EU AI Act is the difference between the provider’s responsibility to support technical marking and the deployer’s responsibility to make certain public disclosures.

Under Article 50(4), the deployer of an AI system must disclose when published image, audio, or video content constitutes a deepfake. The Act defines a deepfake as AI-generated or manipulated content that resembles an existing person, object, place, entity, or event and could falsely appear authentic or truthful.

In a typical Sequencer production, the deployer would be the customer using the platform professionally: for example, a studio, agency, brand, broadcaster, or production company.

That means the responsibility for deciding how a final production should be visibly disclosed generally sits with the client publishing or distributing the content, not with Sequencer.

Sequencer does not automatically add visible AI labels to customer productions. The appropriate disclosure depends on the content, the distribution context, the customer’s role, and the legal analysis applicable to the final publication.

This is especially important for creative works. The AI Act states that when generated or manipulated content forms part of an evidently artistic, creative, satirical, fictional, or analogous work, disclosure may be made in an appropriate manner that does not interfere with the display or enjoyment of the work.

Depending on the production, the client might decide to place that disclosure in:

Sequencer’s role is not to make that editorial or legal decision for the customer. Our role is to provide the governance record and production information the customer needs to make and document that decision.

Sequencer is building governance and provenance directly into the generation and production process.

Our goal is to allow professional creative teams to use multiple generative models while retaining a consistent and verifiable record of how their assets were produced.

The Sequencer Compliance Center gives enterprise customers a centralized view of the governance status of their AI productions.

Rather than forcing a studio or agency to reconstruct an asset’s history after it has been exported, Sequencer records relevant production information as the work is created.

Depending on the production, the Compliance Center can track:

This creates a consistent governance layer even when a production involves multiple models, vendors, or creative processes.

AI content is rarely produced in a single step.

An image might be created with one model, animated using another, modified through generative editing, upscaled, composited into a larger scene, color-corrected, and exported into several versions for different channels.

Sequencer’s governance tracking preserves the history of those transformations.

Each asset can be connected to a record showing:

This enables enterprise teams to move beyond a generic statement that “AI was used.”

Instead, they can maintain a more precise production history explaining what was generated, what was edited, which elements remained unchanged, and which tools contributed to the final result.

That information can support internal governance, client approvals, rights management, legal review, distribution decisions, and responses to future provenance questions.

Content involving real people presents different legal and reputational risks from content depicting fictional characters, products, or environments.

For enterprise customers, Sequencer includes a dedicated real-person workflow.

When a production includes a recognizable individual or uses a real person’s face, voice, or likeness, the asset can be routed through additional governance and documentation steps.

The workflow allows enterprise teams to track:

This process does not replace the customer’s legal review. It ensures that content involving real people does not pass through the same production workflow as an ordinary fictional asset without additional scrutiny.

The final decision about publication, labelling, and disclosure remains with the client deploying and distributing the content.

Ordinary metadata can be lost.

A file may be downloaded, transcoded, compressed, edited in another application, uploaded to a distribution platform, or passed between multiple production partners. Each of those actions can remove or alter information embedded in the original file.

Sequencer therefore uses cryptographic linkages to connect generated assets with their corresponding governance records.

The objective is to establish a verifiable relationship between a media asset and the production history recorded within Sequencer.

Depending on the implementation and workflow, this linkage can help authorized teams verify that:

This does not require every prompt, source file, or confidential production detail to be made public.

The technical layer can establish that a trusted governance record exists while allowing the customer to control access to commercially sensitive production information.

Cryptographic linkage is therefore not simply a disclosure mechanism. It is a way of connecting an asset to an auditable chain of production activity.

The EU AI Act assigns different responsibilities to providers and deployers.

Sequencer provides the generation environment, governance tracking, enterprise review workflows, and technical connections between assets and their production records.

The customer remains responsible for evaluating the final use of the content, determining whether it constitutes a deepfake or another type of disclosable material, and applying any visible notice required when that content is published.

This division of responsibility is intentional.

A production platform does not always know how an exported asset will ultimately be used. The same generated video might remain an internal concept, appear in an explicitly fictional commercial, be included in a documentary, or be published in a context where audiences could interpret it as a recording of a real event.

The disclosure analysis depends on that final context.

Sequencer gives customers the information needed to conduct that analysis while maintaining control over their creative, editorial, and legal decisions.

The first generation of AI creative tools focused primarily on outputs:

Enter a prompt. Generate an image. Download a video.

That model is not sufficient for professional production.

Studios need to know which models were used. Agencies need to document how client assets were transformed. Brands need to understand whether a real person appears in a synthetic production. Legal teams need access to rights and consent records. Publishers need to decide whether and how the final content should be disclosed.

The EU AI Act reinforces an important principle: professional AI generation cannot be separated from provenance and governance.

Sequencer is being built around that principle.

We want creative teams to retain the speed and flexibility of generative AI while gaining the operational controls expected from professional production software.

That means bringing content generation, model governance, production history, consent documentation, real-person review, and verifiable asset records into one workflow.

The future of AI production will not be defined only by who can generate the most impressive video.

It will also be defined by who can explain where that video came from, how it was created, which models were involved, who authorized the underlying materials, and whether its history can be verified.

That is the infrastructure Sequencer is building.

This article provides general information and does not constitute legal advice. Organizations should consult qualified counsel regarding the application of the EU AI Act and other laws to their specific products, workflows, and publications.