When an agency delivers campaign assets, the advertiser expects to own them. That expectation is built into the relationship, the pricing, and the way production budgets have worked for decades. But when those assets are produced using AI tools, the ownership picture changes in ways that production agreements may not yet take into account.

This should concern procurement because it affects what you're actually buying. If the creative assets delivered by your agency were generated by AI, you may be paying production rates for work that cannot be legally protected as your intellectual property. A competitor could potentially use similar or identical outputs without consequence, because there may be no enforceable ownership to prevent it.

Earlier this year, BBS spoke with Chris Mammen of Womble Bond Dickinson about the U.S. legal landscape. We followed up with Roch Glowacki, a commercial contracts lawyer at Lewis Silkin LLP who advises advertisers and agencies on AI governance. We asked him where the law currently stands across jurisdictions, because for global advertisers, the answer depends on where your content is being produced and where it appears.

United States: Clear position, no protection

U.S. law requires human authorship for copyright protection. AI-generated content cannot be copyrighted. This has been affirmed by the U.S. Copyright Office and upheld in federal court. In March 2026, the Supreme Court declined to hear a challenge to this position. Chris Mammen covered this in depth in the earlier BBS interview.

United Kingdom: Untested and now uncertain

The UK Copyright, Designs and Patents Act 1988 contains a provision, Section 9(3), that appears to protect "computer-generated works" by granting copyright to "the person who made the necessary arrangements" for the work to be created. For years, practitioners assumed this would cover AI-generated content.

Roch noted that this assumption is now in serious doubt. There is growing consensus that Section 9(3) may not actually grant copyright to the person who prompted the AI tool. This has not been tested in court, and the UK government's consultation on AI and IP has not produced clear direction. Advertisers operating in the UK cannot rely on this provision as settled law.

China: A middle ground based on human creative input

Li v. Liu, Beijing Internet Court, November 2023

A blogger used Stable Diffusion to generate an image using over 150 prompts, carefully arranging their order, setting specific parameters, and iterating through multiple versions. When another blogger used the image without permission, the court ruled that the plaintiff's substantial intellectual effort through extensive prompting and refinement meant the image qualified as a copyrightable work. The court distinguished this from a simple, single-prompt generation. This is the first court decision globally to grant copyright protection to an AI-generated image based on the degree of human creative input.

Roch described this as an intuitively sensible middle ground. The more human effort goes into directing the AI, the stronger the case for protection. But where that line falls remains vague and untested in other jurisdictions.

What this means for production decisions

Not all AI-generated content needs the same level of protection. Roch raised an important practical distinction.

For long-lived brand assets like hero creative, visual identity, and campaign imagery designed to be built upon over multiple seasons, the absence of copyright can leave you exposed. Roch described a situation from his practice where a startup generated its logo using an AI tool, had no registered trademark, and when a competitor copied the logo, Lewis Silkin could not construct a viable claim to stop them.

For ephemeral content designed for short-term social media impact, the calculus is different. In June 2025, prediction market platform Kalshi aired a fully AI-generated commercial during an NBA Finals game, created using Google's Veo 3 at a cost of approximately $2,000 over two days. The ad went viral. Kalshi almost certainly did not care whether anyone reused clips. Maximum distribution was the goal.

The distinction matters for how you evaluate production costs. If an agency is using AI to produce content more efficiently, that should be reflected in what you're paying. And if the resulting work cannot be protected as intellectual property, your production agreement should account for that, both in the pricing and in the representations the agency makes about what it is delivering.

What advertisers should consider

Understand what role AI is playing in the creative work your agencies are producing for you, and whether the outputs are protectable in the markets where they will appear. For long-lived brand assets, the absence of copyright protection creates real commercial risk. For ephemeral social content, the risk profile is different. Both your production costs and your contract terms should reflect which category the work falls into.

Roch Glowacki is a commercial contracts lawyer at Lewis Silkin LLP. BBS Worldwide is an independent advertising production consulting firm.

This article is for informational purposes only and does not constitute legal advice. The legal perspectives shared here reflect general observations about the current state of the law and should not be relied upon as guidance for specific legal decisions. Readers should consult qualified legal counsel for advice on their particular circumstances.