← Back to Insights
You built the campaign. Do you own it?
"The productivity gain from AI creative is real. The ownership gap it creates is not."
$1.5B
Largest copyright settlement in US history: Bartz v. Anthropic, August 2025.
The Fusion Equation
Performance × Responsibility = Value
Performance
Creative Velocity
Responsibility
IP Defensibility
"AI has to be in service of human creativity. They are tools in the hands of creators. They are never meant to replace them. That is the ethos." — Neal Mohan, CEO, YouTube, Irish Times, January 2025
The core tension
The brand that cannot establish human authorship in its AI creative cannot establish copyright. The content library is performance without protection.
The creative velocity advantage and the IP defensibility requirement are not in conflict. The architecture that documents human control at key decision points delivers both.
The analytical depth
First: scale. McKinsey (2025): 78% of organisations now use generative AI regularly, with marketing and content among the fastest-adopting functions. AI-assisted teams produce more variants, faster, at lower unit cost.
Then: the legal doctrine. The US Copyright Office confirmed in January 2025: zero copyright for AI-generated content without documented human authorship. The Supreme Court denied appeal in March 2026. The doctrine is now settled.
Result: the gap. The brand builds a library of AI-generated assets it cannot legally own, cannot enforce against competitors, and cannot transfer at value in an M&A process.
Only 27% of organisations review all AI-generated content before use (McKinsey 2025). The remaining 73% have no audit trail, no authorship documentation, no ownership architecture.
Adobe / Bartz v. Anthropic
Commercial IP Indemnity · Zero Copyright Doctrine · $1.5B Precedent
27%
Organisations that review ALL AI-generated content before use. The remaining 73% deploy AI-generated content with no full review, no audit trail, and no documented authorship. Content without an ownership architecture is performance without legal protection.
(McKinsey State of AI, March 2025, 1,993 respondents)
In August 2025, Anthropic settled Bartz v. Anthropic for $1.5 billion, covering approximately 500,000 works at $3,000 per pirated title. The court found that the decisive threshold was legal sourcing: training data obtained through shadow libraries cannot qualify for a fair use defense, regardless of whether the training use itself might otherwise be transformative. The ruling establishes that every enterprise deploying AI systems trained on improperly sourced data carries latent copyright liability. The question is no longer whether AI training constitutes fair use in the abstract. It is whether your vendor sourced its training data legally and whether you have documentation to prove it.
"Adobe Firefly is the clearest enterprise-grade response to the IP gap. Every output is generated from licensed or public domain content, and enterprise customers receive commercial indemnity for any IP-related claims. The architecture decision Adobe made, training exclusively on legally sourced content, is the only approach that makes the creative performance advantage legally defensible. The brands that have not asked their AI vendors the same question have not yet made the governance decision."
Adobe Firefly is the only enterprise-grade AI creative tool that offers commercial indemnity. The architecture decision Adobe made, training exclusively on legally sourced content, is the only approach that makes the creative performance advantage legally defensible.
Performance
Creative Velocity
Generative AI has fundamentally changed the economics of creative production. Marketing teams that have integrated AI creative tools report lower unit costs, faster production cycles, and the ability to test creative variants at a scale previously reserved for the largest budgets. McKinsey (2025): 78% of organisations now use generative AI in at least one function, with marketing and product development among the earliest and deepest adopters. The competitive advantage is real, measurable, and compounding for brands that have built effective creative workflows around AI.
Responsibility
IP Defensibility
The copyright question is not abstract. It is a P&L question. A brand that cannot establish copyright in its most commercially successful campaigns cannot protect them from copying, cannot enforce exclusivity in market, and cannot transfer them at value in an M&A process. The architecture that preserves IP defensibility, documented human authorship at key creative decision points, legally sourced training data, clean vendor provenance, does not eliminate the AI creative advantage. It is the condition under which the advantage can be held, transferred, and compounded over time.
Download the full case
PDF · 7 slides · Free access · Downloaded 0 times
Unresolved tensions
Does the EU AI Act change the copyright equation for AI creative in Europe?
When does the 'sufficient human authorship' threshold become commercially unworkable?
How does AI-generated creative affect brand equity valuation in M&A due diligence?
By Fabrice Macarty
This case resonates?
Your brand IP is worth what you can defend. Document human authorship. Source AI training data legally. Build the architecture before M&A reveals the gap.
Start the conversation