The talent and influencer marketing industry has an AI rights problem. Not because agencies are resisting artificial intelligence, but because too many still see rights management as someone else’s responsibility.
The mindset reflects a longstanding weakness in the creative supply chain: its reliance on layers of trust. Brands trusted agencies. Agencies trusted production companies. Production companies trusted talent representatives. Each party relied on assurances that rights had been cleared further up the chain. The pressure on that model is now becoming clear, as Benjamin Woollams (pictured), Founder and CEO, TrueRights, explains…
That model no longer works
AI has fundamentally changed the economics of creative production. A campaign is no longer a handful of carefully managed assets. It might be hundreds or thousands of synthetic images, cloned voices, localised edits and AI-generated variations produced in days rather than months.
At that scale, rights management stops being a legal exercise at the end of production and becomes part of the production process itself. The problem is that rights are still generally recorded in contracts, rather than linked to assets in a form that production teams can check. A PDF in a contract folder, for example, isn’t a rights management system. It’s a document that nobody checks when an editor creates a new version for social media, when an AI model generates another variation or when an asset is reused six months later in another territory. By then, usage has already drifted beyond the original licence, leaving the agency exposed.
“It’s not our problem” isn’t a rights strategy
Agency sign-off does not, by itself, settle the questions that matter. Did the person approving the work actually have authority? Did the permission include synthetic regeneration? Did it extend to AI training, localisation or future markets? Was the licence limited by territory, category or duration? The recent dispute involving Dua Lipa and Samsung illustrates exactly why those distinctions matter.
Samsung says it relied on assurances from a content partner that all necessary permissions had been secured, including for retail packaging. Dua Lipa alleges that those rights were never granted, and the case remains unresolved. The point for agencies is that an assurance is not a verifiable rights record. If the underlying permission cannot be checked, every organisation further down the chain inherits the exposure. The photograph at the centre of the dispute wasn’t even AI-generated. If the industry struggles to evidence ownership and licensing for a single static image, how will it cope with thousands of synthetic campaign assets, each with its own derivatives and usage rights? This is why agencies can no longer afford to think, “it’s not our problem.”
Creating confidence to use AI at scale
Unlicensed capability is a commercial risk for any agency using generative technology. The ability to evidence permission, provenance and chain of title will become a competitive advantage.
That requires rights to become live data rather than static paperwork. Every asset should carry structured information showing who owns it, how it can be used, where it can appear, when permissions expire and whether AI training or synthetic regeneration is permitted. Provenance, watermarking and fingerprinting should be applied at the point of creation, although none replaces a properly scoped license or record.
Most importantly, rights management cannot stop when a campaign launches. AI assets need continuous monitoring, automated renewals, enforcement mechanisms and payment workflows that evolve alongside the content itself. When done well, this gives agencies and clients the confidence to use AI at scale. Agencies that continue relying on verbal assurances, PDF contracts and historic workflows risk more disputes and a gradual loss of their place in the value chain.
Conclusion
Brands are already asking harder questions about provenance, compliance and chain of title. Purpose-built right management infrastructure is beginning to provide those answers directly into their platforms, linking permissions, provenance, and chain of title data throughout their lifecycle. Agencies that cannot provide that confidence will find themselves being bypassed by clients.
The opportunity is still there. Agencies have always existed to reduce complexity between brands, talent and production. In an AI economy, that role becomes even more valuable when proper rights management is built into the infrastructure of production.
