The age of AI – Hong Kong plans consultation on copyright protection

The Director of the Hong Kong Intellectual Property Department, David Wong Fuk-loi, has recently told online service providers that the administration intends to carry out a consultation on modernising Hong Kong’s copyright laws to take account of the rapid development of artificial intelligence (AI).

Mr Wong was quoted as saying, “in response to the prevalence of AI, we are to release a consultation paper within this year and review what other areas in the current Ordinance need to be amended.” On a separate occasion, the Secretary for Innovation, Technology and Industry suggested that the consultation will "explore further enhancement of the relevant protections provided by the Copyright Ordinance so as to ensure that Hong Kong's copyright regime remains robust and competitive". The proposed consultation was first trailed in the Chief Executive’s 2023 Policy Address.

Recent amendments to the Copyright Ordinance

The planned consultation comes less than three years since a previous consultation which took place from November 2021 to February 2022. At the time, a government spokesperson said that Hong Kong’s copyright regime was “over a decade behind international developments”. In 2021, the government expressed the ambition for Hong Kong to turn itself into a regional IP-trading centre.1 This ambition was re-stated in the 2024-2025 budget presented by Financial Secretary Paul Chan on 28 February 2024.2

Following the earlier consultation, the government announced significant and long-awaited amendments which were incorporated into the Copyright Ordinance on 1 May 2023:

• give copyright owners a technology-neutral exclusive communication right so as to ensure that the protection afforded to them would cover their works communicated to the public through any mode of electronic transmission;
• introduce criminal sanctions against infringements of the exclusive communication right;
• provide for new fair dealing exceptions for the use of copyright works for three categories of purpose: parody, satire, caricature and pastiche; commenting on current events; and quotation of copyright works;
• a “safe harbour” regime to limit the liability of online service providers (OSPs) for copyright infringement acts by their subscribers on their service platforms, provided that OSPs meet certain prescribed conditions; and
• facilitate the use of copyright material in online learning and in libraries, museums and archives.

The amendments were intended to ensure that copyright owners benefit from copyright protection when their works are communicated through digital forms of transmission (such as online streaming) while balancing the interests of users with new exceptions. 

Following the 2021/2022 consultation, the government indicated that it would first address the most imminent and fundamental issues and will look at other new and emerging copyright issues such as AI/copyright in the future.

The current copyright position in Hong Kong

Under the Copyright Ordinance, the author of a work is generally the first owner of the copyright in it. Further, a work qualifies for copyright protection only if the author is a natural or legal person. This rules out an AI tool claiming copyright ownership.

In order to benefit from copyright protection, the work must be “original”. In Hong Kong, the threshold for originality is a low one – a work need only be “original” in the sense that the author created the work by their own efforts (as opposed to slavishly coping from another work) and expended more than negligible or trivial effort or relevant skill in the creation of the work. A work may be original even if the author has drawn on existing knowledge or material.

If third parties do acts such as copying or reproduction without the consent of the copyright owner, they may be subject to a claim for copyright infringement.

However the rapid development of AI has left some crucial questions unanswered such as:

• Is AI-generated content capable of satisfying the test of “originality” to attract copyright protection?
• Whether and if so, how much human input is required during the generation in order to benefit from copyright protection?
• Who should be the copyright owner?
• To what extent can AI developers use third party content to train AI without infringing copyright?

In Hong Kong, it is arguable that AI-generated works already attract copyright protection based on the existing provision in the Copyright Ordinance which applies to a work generated by a computer in circumstances where there is no human author of the work. Section 11(3) provides that for “a literary, dramatic, musical or artistic work which is computer-generated, the author is taken to be the person by whom the arrangements necessary for the creation of the work are undertaken.” In the context of AI-generated works, one uncertainty here is who will be considered as the person making the “arrangements necessary for the creation of the work”.  For example, would this be the AI developer or the users feeding prompts to generate the work?   

Different jurisdictions are approaching these issues in different ways.

United Kingdom

The UK government published its response to the AI white paper A Pro-Innovation Approach to AI Regulation on 6 February 2024. The response says there will no new legislation for AI in the UK with the emphasis on pursuing an agile sector-based approach. The government has abandoned plans to try to reach agreement with stakeholders on a new AI voluntary code of practice, which would have set out rules on the training of AI models using copyrighted materials and which had been intended as a compromise between the wish of AI developers to access good quality data to train their large language models (LLMs) on the one hand and the interests of content creators wishing to maintain control over their copyrighted work on the other.

The government has now said it is exploring ways to ensure better transparency in the information provided to rightsholders so they can know whether content they have produced has been used as an input for AI modelling. The government says it will “lead a period of engagement with the AI and rights holders sectors, seeking to ensure the workability and effectiveness of an approach that allows the AI and creative sectors to grow together in partnership”.

A few days earlier, the House of Lords Digital and Communications Committee published a report on generative AI and LLMs which said that the current legal copyright framework was failing to ensure that creators are properly rewarded for their work and prevent others from using their work without permission.3 The report urged the government to publish its view on whether copyright law does provide sufficient protection for rightsholders.

Copyright owners had previously been encouraged by the decision not to expand the current exception for “text and data mining” (TDM) to allow for the mining of all copyright material with no opt-out right available for the rights holders. At present, the TDM exception only encompasses non-commercial research by those who enjoy lawful access to the copyright works.

United States

In September 2023, the US Copyright Office issued a “Notice of Inquiry on Copyright and Artificial Intelligence”. The notice sought public input regarding “the scope and level of human authorship, if any, in copyright claims for material produced in whole or in part by generative AI”.

Amongst specific measures being explored, were the availability of technical tools that could help in labelling AI outputs and whether outputs that mimic the style of particular artists or performers, should be banned.

US copyright law does not allow for copyright protection for works generated purely by generative AI. Where generative AI tools have been used to create a piece of content, copyright protection will extend to those parts with direct human authorship and not the parts generated by generative AI.

A presidential directive issued in October 2023 requires the Under Secretary of Commerce for Intellectual Property and Director of the US Patent and Trademark Office to issue guidance by July 2024 on issues such as the scope of protection that should be available for AI-generated works and how copyrighted works used in AI training should be protected.

Tricky questions

In light of the above, some issues the Hong Kong consultation may want to address include:

• Would a purely AI-generated work be protected by copyright law?
• If not, what degree of human intervention/input will be necessary for a work to benefit from copyright protection?
• Insofar as an AI-generated work can be viewed as a copyright work, who would own the copyright in it?
• Whether and how does the current “computer-generated works” statutory provision apply to an AI-generated work?
• How might an AI developer satisfy themselves that their use of training data does not infringe third party copyright works? Will specific exceptions such as text and data mining and further fair dealing exceptions be introduced in the Copyright Ordinance? Does the current “research” exception cover any types of machine-learning?
• How would a copyright owner establish that an AI-generated work infringes the copyright in their original work?
• If an AI-generated work is infringing, who should be liable for the copyright infringement? The AI developer, the user or some intermediary?

Until specific laws and official guidelines are in place, businesses deploying AI are advised to have in place the right contracts and policies to address the IP risks presented by generative AI. The Hogan Lovells Intellectual Property Team is well placed to assist.