Login / Register

Market OpportunitiesTech & InnovationEntrepreneurshipSustainabilityCreativity & LifestyleHKTDC Updates

My Feed

Tech & Innovation

AI presents IP challenge

Information Techn...RoboticsAIDesignArt

A licensing lawyer points out that key jurisdictions deny copyright to machine-generated content.


Generative AI is drawing intense attention in the licensing industry, as it could cost-effectively generate vast amounts of IP content, but comes with a range of legal, privacy and security problems.

Underlining the importance of the issue, the discussion session Generative AI: Navigating Intellectual Property Risks and Challenges at the recent Asian Licensing Conference (ALC), organised by the Hong Kong Trade Development Council (HKTDC, drew a full house.

A key legal question is whether AI-generated content is, in fact, IP. If it is not, then it cannot be licensed.

Alan Chiu, Managing Partner at legal practice Ellalan, told the audience that this had become a complicated issue, with courts in two leading IP markets, Mainland China and the United States, making opposite rulings.

In one high-profile US case, computer scientist Stephen Thaler created an image-generating computer system ,which produced a work A Recent Entrance to Paradise in 2012. The US Copyright Office declined to register the work, a decision upheld by the District Court for the District of Columbia. The courts held that the work by the Creativity Machine was not made by a human, so could not be copyrighted. The computer scientist will appeal.

In another high-profile American case, artist Jason Allen, who won a fine art competition with his AI-generated work Theatre d’Opera Spatial, also had his application for copyright turned down.

The artist pointed out that he had made more than 600 prompts into the AI system in order to generate the work, but the Copyright Office said, with a human creator absent, copyright was not possible.

In Mainland China, an artist created an image of a woman using more than 400 prompts on the generative system Stable Diffusion. A blogger lifted and used the image, and the artist, surnamed Li, successfully sued. The court recognised the effort in directing the system to create the AI work and awarded damages.

These differing views led an attendee at the session to pose the question: “Was it best to avoid AI-generation, if work cannot be copyrighted?”

Mr Liu said AI is now where the Internet had been 30 years ago, with many potential users still sceptical.

He predicted that the questions would soon be resolved, so the licensing industry needed to prepare. Companies needed to lay down policy guidelines to create transparency for themselves and clients.

He pointed out that, if a user prompted an AI system to produce a mouse cartoon, the end result may resemble Mickey Mouse, but the prompter had not intended this, so there would not be infringement.

Mr Liu recommended that licensees and licensors include AI clauses in their agreements to offer indemnity and govern the approval process.

He said training in AI use might be necessary, as laws are being introduced to govern the technology. For example, the European Union is enacting legislation this month laying down general guidance on AI use, ensuring applications do not infringe on basic human rights or violate existing copyright law.

The session moderator Roger Berman, President of ZenWorks, said generative AI presents a moving target for the licensing industry and is rapidly evolving.

Discover More

Subscribe free e-newsletter

Latest on Asia business intelligence
Mobile-friendly design to read on the go
Customise your news by registering online

Learn about latest market updates and insights to empower your business.