Allen Wang Speaks at LESI 2026 Dublin: Licensing Challenges in the Evolving Film Industry

Against the backdrop of rapid transformation in the global film industry and the deep integration of artificial intelligence into content creation and distribution, copyright licensing and business models are being fundamentally reshaped. 

In April 2026, the LES International (LESI) Annual Conference was held in Dublin, bringing together leading experts from the legal and entertainment industries worldwide. 

Allen Wang, leading Partner at Beijing TA Law Firm, was invited to speak at the workshop “Reel Challenges: Licensing in the Evolving Film Industry,” where he joined international panelists to explore emerging issues in AI-driven content production, cross-border licensing, and the evolving structure of rights and value in the film sector.

Below is a transcript of Allen Wang's key remarks from the LESI 2026 Dublin conference.

Thank you, Anna. I’ll share a few observations from the China perspective, and I’ll try to frame them around how AI and market developments are reshaping licensing structures in practice.

In China today, when we talk about film or audiovisual content, we are no longer talking about a single work. We are increasingly talking about an ecosystem. Film IP is not limited to theatrical or streaming value. It is being developed across multiple downstream scenarios — merchandising, consumer products, collectible cards, blind boxes, and increasingly cultural tourism and immersive experiences.

A good example is Ne Zha 2, which has generated more than 2 billion USD in global box office and is among the top-grossing films worldwide. But more importantly, the box office is only the starting point. The real value is created after the release, through downstream markets. In China, the collectible card and blind box market alone has grown into a multi-billion-dollar industry, and much of that growth is driven by recognizable film and character IP.

Another example is The Three-Body Problem, which is not only a successful Chinese sci-fi IP, but also a globally recognized one, with a Netflix adaptation and ongoing development across different formats. At the same time, it is expanding into VR and AR environments, virtual communities, and location-based experiences. What we see here is not just expansion, but convergence across industries.

This creates a fundamental shift for licensing. Traditionally, licensing focused on the work itself — a film, a distribution right, a territory. But in this new environment, most of the value is generated later, across different industries and use scenarios. So the key issue is no longer simply ownership. It becomes how to structure licensing in advance, in a way that can accommodate future expansion.

In practice, this means licensing discussions are happening much earlier. Companies are engaging with potential partners — including brand collaborations, offline experience operators, and tourism projects — already at the development stage. So licensing is no longer a post-production transaction. It becomes part of the content design process itself. From a legal perspective, that requires much more structured and forward-looking contract design, especially in defining application scenarios, boundaries of use, and alignment with actual business models.

If this is how IP is evolving, then AI is changing something even more fundamental — how content itself is created.

In China, AI is already embedded in production workflows. It is used in animation, in AI-generated live-action styles, in visual effects and post-production, and increasingly in script development and pre-visualization. This is no longer experimental. It is part of real production.

Production cycles are shorter, costs are much lower — especially in post-production — and some creative roles are being restructured. 

But at the same time, this creates an important counter-effect. The more powerful AI becomes, the more valuable strong IP becomes. Because while AI can generate content at scale, it cannot easily replace distinctive ideas, recognizable characters, or culturally meaningful narratives.

So in the AI era, real value still lies in high-quality, identifiable content. AI lowers the barrier to production, but it raises the standard for value.

At the same time, AI is also enabling new content formats. In China, we are seeing the rapid rise of microdramas — short-form, highly serialized, fast-produced, and deeply integrated with platform distribution. These formats are extremely scalable, and AI plays a key role in enabling that scalability. What is particularly interesting is that this model is now expanding globally. So China is not only producing content, but also exporting a new content production model.

Once production changes, legal questions also shift. Traditionally, licensing was built around ownership and chain of title. But now we are facing questions that are much more difficult: who is the author, what is copyrightable, and what exactly is being licensed?

From what we are seeing in practice, the key is not whether AI is involved, but whether human contribution can be demonstrated. That includes control over the process, creative decision-making, and a clear causal link between human input and the final output. Authorship is no longer something we can assume. It becomes something that needs to be demonstrated.

This leads to another very practical issue in China, which is AI labeling. From a regulatory perspective, labeling is intended to promote transparency and content governance. 

But in practice, it has very direct commercial consequences. On some platforms, especially music platforms, once content is labeled as AI-generated, it may not qualify for revenue sharing, and its licensing value may be significantly reduced.

So labeling is no longer a neutral compliance tool. It can become a gatekeeper of economic value. And this raises deeper questions about how decisions are made, what standards are applied, and how creators can respond. At the same time, when content moves across borders, different jurisdictions may treat AI-generated content differently, which further complicates licensing.

Another area where the impact of AI is becoming very visible is in performers’ rights. AI is now capable of replicating likeness, voice, and performance. In some cases, it is beginning to replace actors, particularly those outside the very top tier. So this is no longer only a copyright issue. It becomes a livelihood issue.

At the same time, the risks are not limited to economic substitution. AI-generated use of likeness and voice can also lead to reputational harm, especially when used in inappropriate or harmful contexts. In those situations, the impact on the individual can be significant and difficult to reverse. So the legal issues here extend beyond copyright into personality rights, performers’ rights, and control over downstream use.

In this context, licensing becomes more complex. It is no longer just about granting permission. It also needs to define boundaries — what is allowed, what is prohibited, and how future use is controlled. This is particularly challenging when AI training is involved, because the use of existing content may extend far beyond the original context.

So after looking at all these changes — in IP, in production, and in rights — the real question becomes: where is the actual legal demand?

We are in a situation where AI is widely used, production is becoming more efficient, and content is increasing rapidly. But at the same time, monetization is not always clear. If content is generated more easily, who is the buyer, and how much are they willing to pay? Will lower production costs translate into sustainable value, and will there still be sufficient room for structured legal service fee?

My view is that legal demand is not decreasing. It is shifting. It is moving from reactive protection toward structural design. It is no longer only about enforcing rights after the fact, but about designing how value is created, structured, and allocated in advance.

This includes designing licensing structures for multi-layer IP systems, managing AI-related risks, defining control over outputs and allocation of value, and navigating cross-border differences. Increasingly, it also includes managing reputational and cultural risks, especially in a content industry that does not only create products, but also shapes narratives and public perception.

So in the end, the question is not whether AI creates value. It is who controls that value, and how it is structured.

Thank you for your attention, and I look forward to further discussions with you.