By Lili Kazemi
Executive Summary
- From 2023 through early 2025, generative AI triggered a wave of copyright litigation across music, publishing, film, and visual arts.
- More recently, licensing and negotiated frameworks have begun to emerge alongside litigation as practical tools for managing AI-related IP risk.
- Deals such as Warner–Suno and Disney–OpenAI reflect how rights holders are experimenting with licensing as a way to define acceptable use.
- These arrangements do not replace enforcement or regulation, but they do offer early signals about how value, control, and responsibility may be allocated in the AI economy.
- For general counsel, these developments highlight the growing importance of licensing structure, governance, and downstream tax and transfer pricing implications.
Two things can be true at once in AI right now.
Rights holders are suing AI companies for mass infringement.
Those same rights holders are entering into large-scale licensing agreements with AI platforms.
Warner Music Group’s settlement with Suno and Disney’s licensing and investment agreement with OpenAI are not contradictions. They are markers of a market searching for workable solutions while the law continues to evolve.
This piece builds on questions I explored in When AI Works, Who Gets Paid? and When AI Creates Value, Settlements Aren’t Arm’s-Length — They’re Arms-Twist. What we are seeing is not the end of copyright conflict, but a period where licensing, litigation, and governance are developing in parallel.
Warner–Suno: From Complaint to Contract
The Warner–Suno dispute followed a now familiar script in AI music litigation.
Warner and its affiliated publishers alleged that Suno trained music-generation models on copyrighted sound recordings and compositions without permission, generated outputs substantially similar to protected works, and used that unlicensed training to build a commercial product.
Suno denied the claims, arguing that its outputs did not reproduce copyrighted songs, that its training relied on licensed or user-provided data, and that model training constituted non-expressive, intermediate copying rather than infringement.
What ultimately mattered was procedural posture. When the case settled, the parties were approaching expert discovery, the phase where training datasets, model architectures, and similarity analyses become central, intrusive, and expensive.
Instead of forcing a court to decide whether training on copyrighted recordings is fair use, Warner and Suno converted litigation risk into a licensing framework. The agreement, announced in late November 2025, contemplates fully licensed models beginning in 2026 and—in a surprising strategic move—includes Suno’s acquisition of the live-music discovery platform Songkick from WMG. This suggests the partnership will go beyond training data to include artist opt-in controls for names and likenesses, tiered user access, and a shared upside economic structure that bridges digital creation and live-fan engagement.
Legally, the settlement resolved little. Practically, it defined what acceptable AI music generation might look like in a commercial setting.
Disney–OpenAI: The Magic AI Kingdom
Disney’s deal with OpenAI makes the broader shift easier to see. After escalating enforcement activity in the AI space, Disney announced a one-billion-dollar equity investment in OpenAI on December 11, 2025, and entered into a three-year licensing agreement covering more than two hundred Disney-owned characters for use in Sora and ChatGPT image generation. OpenAI released Sora 2 in October, unleashing a flood of copyright violations onto the internet. But instead of suing the company, or even threatening to do so, Disney has entered into this landmark deal.
The official OpenAI announcement emphasizes that the agreement is deliberately constrained. It excludes real actors’ voices and likenesses, limits use cases, and embeds safety and usage controls intended to protect creators and brand integrity. Notably, while the licensing partnership spans three years, reports indicate OpenAI secured only a one-year exclusivity window. After 2026, Disney retains the flexibility to license its iconic IP to other AI developers, a move CEO Bob Iger has described as preparing for “a way into” on AI partnerships.
That structure matters. While the partnership spans multiple years, exclusivity lasts only one year. This reflects a careful balance. Disney secured early strategic alignment with a leading AI platform, while preserving flexibility to re-license and re-price its IP as the market evolves.
The sequencing also matters. Innovation proceeded amid uncertainty. Negotiation followed once the economic value became clear. This act-first, negotiate-later approach is emerging as a practical response to doctrinal lag, not a statement about legal entitlement.
Licensing Instead of Fighting
What we are seeing is not simply litigation followed by settlement. It is a broader effort to use licensing as a way to frame acceptable behavior in a rapidly changing technological environment.
AI copyright disputes are fact-intensive and technically complex. Courts are being asked to resolve questions about training data, model behavior, and output similarity that do not fit neatly into existing legal categories.
Licensing agreements and private mediation offer flexibility that litigation often cannot. They allow parties to define scope, impose controls, allocate risk, and revisit terms as technology changes.
That does not mean litigation is disappearing. It means licensing is being used alongside enforcement as a tool to manage uncertainty.
The Other Side of the Strategy: Midjourney and Ongoing Enforcement
Disney is not abandoning enforcement. It is using it strategically.
Alongside licensing, Disney and other major studios have pursued claims against AI image generators such as Midjourney, alleging the generation of near-replica depictions of iconic characters. On December 15, 2025, the parties moved toward a resolution by filing a joint selection of the Hon. Suzanne H. Segal, a retired magistrate judge with extensive settlement experience, to serve as the private mediator for the consolidated Disney and Warner Bros. cases.
At the same time, enforcement activity continues elsewhere. Just days before the OpenAI deal, Disney sent a massive cease-and-desist letter to Google alleging large-scale infringement via its Veo video tool. By December 15th, Google had already begun wiping dozens of infringing videos from YouTube, with links now displaying a standard copyright claim notice from Disney. Class actions against companies such as Adobe and others continue to challenge how training data has been sourced and used.
Copyright litigation is very much alive, even as licensing gains traction.
Who Gets Paid? The Value Chain Revisited
When I asked When AI Works, Who Gets Paid?, I framed AI as a value chain that includes IP owners, data contributors, model developers, infrastructure providers, distributors, and end users.
The Warner–Suno and Disney–OpenAI deals show how that chain is being rebalanced.
IP owners are moving from plaintiffs to counterparties. Warner secured a forward-looking revenue interest. Disney is monetizing authorized access to one of the most advanced multimodal platforms in the market.
AI platforms, meanwhile, are discovering that legitimacy and scale increasingly depend on access to premium IP, and that access comes with governance obligations and shared economics.
This is what a market adjustment looks like.
Five Things General Counsel Should Be Watching
1. Licensing is becoming a governance layer
Licensing now sets norms for use, embeds controls, and defines acceptable behavior while the law remains unsettled.
2. Exclusivity is shorter and more strategic
Time-limited exclusivity captures early leverage without long-term lock-in, preserving future optionality.
3. Litigation remains a core feature of the landscape
Training-data claims, output-based theories, and class actions continue. Licensing does not eliminate enforcement risk.
4. Regulatory attention is increasing
Federal and executive-level focus on AI safety and accountability suggests private deals will eventually intersect with clearer regulatory expectations.
5. Tax and transfer pricing consequences follow structure
Licensing terms affect withholding, characterization, permanent establishment risk, and profit allocation. Transfer pricing analysis will increasingly hinge on control, risk, and governance, not just headline pricing.
Final Takeaway
In the AI economy, the market may not be speaking first, but it is speaking more clearly and more insistently. The signal is getting louder.
Licensing and IP governance are becoming central to how AI systems are built and constrained. Courts and regulators remain essential, but negotiated frameworks are increasingly shaping day-to-day practice.
The challenge ahead is not choosing between markets and law, but learning how to listen to both at the same time.
Read More
When AI Creates Value, Settlements Aren’t Arm’s-Length — They’re Arms-Twist
Disney, Midjourney select Hon. Suzanne H. Segal as private neutral for mediation
Warner Music Group and Suno Forge Groundbreaking Partnership
Google Wipes Disney AI Videos From YouTube Following Legal Threats
Lili Kazemi is General Counsel and AI Policy Leader at Anant Corporation, where she advises on the intersection of global law, tax, and emerging technology. She brings over 20 years of combined experience from leading roles in Big Law and Big Four firms, with a deep background in international tax, regulatory strategy, and cross-border legal frameworks. Lili is also the founder of DAOFitLife, a wellness and performance platform for high-achieving professionals navigating demanding careers.
Follow Lili on LinkedIn and X
🔍 Discover What We’re All About
At Anant, we help forward-thinking teams unlock the power of AI—safely, strategically, and at scale.
From legal to finance, our experts guide you in building workflows that act, automate, and aggregate—without losing the human edge.
Let’s turn emerging tech into your next competitive advantage.
Follow us on LinkedIn
👇 Subscribe to our weekly newsletter, the Human Edge of AI, to get AI from a legal, policy, and human lens.
Subscribe on LinkedIn
