In the AI era, intellectual property protection is not merely a matter of legal rules. It is also an infrastructure question: can rights be identified, recorded, licensed, traced, and enforced in practice?
Over the past period, much of the discussion around AI and intellectual property has centered on several questions: Are AI-generated works copyrightable? Does the use of training data amount to infringement? Should platforms and model developers bear liability?
But once AI is genuinely embedded in enterprise workflows for content production, brand communications, advertising, product design, and software development, a more fundamental issue begins to surface.
Even if the law recognizes a right, can an enterprise prove it? Can a platform identify it? Can a system trace it? And when a dispute arises, can the evidentiary chain support the enterprise in bringing or defending a claim?
That is why I have recently been paying attention to WIPO's launch of AIII. AIII is a reminder that intellectual property protection in the AI era must be understood anew: it is not only a matter of rules, but also a matter of infrastructure.
For intellectual property service providers, the value of copyright services is likewise shifting from "helping companies obtain certificates" to "helping companies build evidentiary systems capable of supporting future disputes."
01 What Is AIII
AIII is not merely about copyright ownership; WIPO is building a more foundational framework for dialogue
On March 17, 2026, the World Intellectual Property Organization (WIPO) launched the Artificial Intelligence Infrastructure Interchange, or AIII, in Geneva. WIPO pronounces the acronym "A-triple-I."
The name may be understood as a mechanism for dialogue on AI infrastructure. In the intellectual property context, it can also be understood as a mechanism for dialogue on AI and IP infrastructure.
According to WIPO's official description, AIII is not a mechanism for directly making legal rules, nor is it a forum devoted simply to debating whether AI-generated works should enjoy copyright protection. Instead, it focuses on the technical and operational issues of the intellectual property system in the AI context and seeks to foster dialogue among creators, right holders, AI developers, technology companies, academia, and public institutions.
AIII is concerned not with the abstract question of "who should own the rights," but with how rights can be identified, recorded, licensed, traced, and enforced.
This is precisely the critical shift in intellectual property issues in the AI era.
02 Why Discuss Infrastructure Now?
In the past, intellectual property protection could often rely on registration, contracts, manual review, platform takedown procedures, and ex post enforcement. Imperfect as that system was, it could still function in an era when content was produced at a relatively limited pace and dissemination chains were comparatively clear.
Generative AI has changed that premise. Content can be produced at scale, source materials can be rapidly remixed, works can circulate across platforms, model training may involve massive datasets, and AI outputs can in turn appear in advertising, short-form video, e-commerce pages, product packaging, software code, and product design.
Rights issues are no longer confined to a single work, a single contract, or a single act of infringement. They have become systemic problems that arise continuously, spread rapidly, and are difficult to trace back.
WIPO's focus in AIII on watermarking, fingerprinting, rights management, and standards for identification and attribution is not because these concepts are new; it is because the AI era requires an infrastructure through which machines, platforms, enterprises, and courts can all recognize rights.
AI has amplified content-production capacity, but it has also amplified the difficulty of proving rights.
03 Implications for Enterprises
Many companies are still discussing AI compliance at the tool level
Many companies still approach AI compliance at the level of tools: Which AI tool was used? Was it paid for? Is commercial use permitted? Should a disclaimer be added on the page?
Yet the real risk often lies elsewhere: whether the process by which content came into being can be clearly explained.
- Where did the source materials come from?
- Who participated in the creative process?
- What did the AI do, and what did the human do?
- Was authorization obtained?
- How did the versions evolve?
- What was the final scope of commercial use?
If a promotional image, a piece of advertising copy, a product description, or a design proposal is later accused of infringement, the company cannot simply answer, "It was AI-generated." It must also be able to answer all of the questions above.
Compliance for AI-generated content is not about attaching a label to the result; it is about preserving evidence of the process.
Truly valuable compliance does not mean using less AI. It means retaining the materials that will later enable the enterprise to explain, prove, and defend its conduct while using AI.
04 Service Upgrade
From copyright registration to an evidentiary system
In the past, when companies spoke of copyright protection, their first instinct was often to pursue copyright registration. Registration remains important, of course. It is one part of proving rights, and it can play a role in platform complaints, licensing transactions, enforcement negotiations, and litigation.
Today, however, registration alone is no longer sufficient to cover the complexity of the scenarios companies face. This is especially true when content results from employee authorship, outsourced design, collaborative development, licensed materials, AI-assisted generation, and multiple rounds of revision. A single registration certificate can prove one result, but it cannot necessarily explain how that result was formed.
- Traceability of the creative process and version history;
- Provenance of source materials and the chain of authorization;
- Points at which AI was used and the boundary of human involvement;
- Records of publication and commercial exploitation;
What truly matters going forward is helping enterprises build a comprehensive copyright evidence-preservation mechanism. The value of specialized copyright counsel is shifting from "helping companies obtain certificates" to "helping companies build evidentiary systems capable of supporting future disputes."
05 Asset Governance
Copyright protection must become part of enterprise content-asset governance
If copyright services in the past were centered more on registration, contracts, and enforcement, then in the AI era specialized copyright counsel needs to enter the content-production process much earlier. Its role is not merely to handle disputes, but to help enterprises design the evidentiary structure before disputes arise.
A company's marketing, branding, design, research and development, and product teams are all generating content, but that content is often dispersed across different personnel, suppliers, tools, and platforms. Where counsel can truly add value is by turning these fragmented outputs into manageable copyright assets.
- Which content genuinely belongs to the company?
- Which content is only licensed for use?
- Which content is subject to sublicensing restrictions?
- Which content involved the use of AI?
- Which content is suitable for registration?
- Which content requires enhanced evidentiary recordkeeping?
In this way, copyright legal services cease to be merely something to call on after a problem arises and instead become part of enterprise content-asset governance. Copyright services of the future will not be limited to handling infringement; they will help companies turn content into assets that can be proven, licensed, and enforced.
06 A System
Registration is the starting point; the system is the moat
WIPO's launch of AIII in fact sends a signal: intellectual property protection in the AI era cannot rely solely on isolated rules or a single certificate. If rights are to function in practice, they must be embedded in a system capable of identification, recordation, licensing, traceability, and enforcement.
For enterprises, copyright registration remains important, but it is only the starting point. What will truly protect a company in future disputes is an evidentiary system built around the creative process, source materials, chain of authorization, use of AI, version evolution, and commercial publication.
Registration is a certificate; evidence preservation is a system.
In the AI era, copyright protection of real value is not merely about proving what one once owned. It is about enabling enterprises to explain clearly, produce evidence, and prevail in future disputes.
In the past, we were more accustomed to treating intellectual property as a form of ex post protection. If infringement occurred, we enforced; if proof was needed, we registered; if a dispute arose, we supplemented the record afterward.
In the AI era, that cadence is being disrupted. Content is produced faster, distributed faster, remixed faster, and disputes arise earlier. For enterprises, what now truly matters is no longer whether there is a certificate at a single point in time, but whether the entire process by which content came into being has left behind an evidentiary structure that can be identified, explained, and submitted.
That is why I believe AIII's significance for the intellectual property services industry lies not only at the level of international rulemaking, but also at the level of business models. Legal services that will matter in the future are not merely those that complete registrations on behalf of enterprises, but those that help embed copyright protection into content production, commercial use, and asset-governance workflows.
In the AI era, intellectual property protection of real value is not merely about proving what one once owned. It is about enabling enterprises to explain clearly, produce evidence, and prevail in future disputes.