Managing IP is part of the Delinian Group, Delinian Limited, 4 Bouverie Street, London, EC4Y 8AX, Registered in England & Wales, Company number 00954730
Copyright © Delinian Limited and its affiliated companies 2023

Accessibility | Terms of Use | Privacy Policy | Modern Slavery Statement

Future of IP – China: A closer look at AI and trade secret protection

Sponsored by


Jing Zhang and Guang Hou of Liu Shen & Associates discuss AI-related innovations and patent protection

Not all artificial intelligence (AI)-related innovations are suitable for patent protection. Some innovations with improved AI algorithms may not have a specific application field that can fit into the protection of patent, and some AI-related IP rights are not protected by the Patent Law, e.g. the data itself.

In these scenarios, the AI software owner may consider protecting it as a trade secret. Considering the nature of secrecy of trade secrets, the advantage of protecting AI algorithms or softwares in this form is that the ‘idea’ of products can be concealed.

On September 10 2020, the Supreme Court of China announced the judicial interpretation titled “Provisions of the Supreme Court on Several Issues Concerning the Application of Laws in the Trial of Civil Cases of Infringement of Trade Secrets” (hereinafter referred to as ‘new judicial interpretation’).

‘New judicial interpretation’ clarifies, specifically, the subject of protection, the burden of proof of the right holder, and the forms of act of ‘use’ to be deemed as infringement, for which an analysis is provided below along with related suggestions.

It affirms that software-related contents can be the subject of trade secret protection, and specifies the requirements for product novelty.

Article 1

The structure, raw materials, components, formulas, materials, samples, styles, new plant variety propagation materials, techniques, methods or their steps, algorithms, data, computer programs and related documents related to technology can be determined by the court to constitute technical information referred to under Article 9, clause 4 of the Anti-Unfair Competition Law.

The algorithm, data, computer program and related documents in the above provision are typical elements involved in AI algorithm or software.

Article 4, paragraph 2

The new information formed after sorting, improving and processing the information known to the public meets the provisions of Article 3 of this interpretation, and it shall be deemed that the new information is not known to the public.

This indicates a low level of requirements on product novelty for trade secret related to data and information processing software.

Suggestion: For assets that may not be protected by copyright, such as big data, artificial intelligence algorithms, and databases, protection by trade secret is suitable for shielding the underlying technical scheme.

It emphasises the ‘reasonability‘ in the requirements of confidentiality measures, specifies the consideration of ‘substantially the same’, and eases the right holder’s burden of proof.

Article 6

If one of the following circumstances is sufficient to prevent the disclosure of commercial secrets under normal circumstances, the court will determine that the right holder has taken corresponding confidentiality measures:

“(5) Taking measures such as prohibiting or restricting the use, access, storage and copying of computer equipment, electronic equipment, network equipment, storage equipment and software that can access and obtain trade secrets;

(6) Require the retired employees to register, return, clear and destroy the trade secrets and their carriers they have contacted or obtained, and continue to undertake the obligation of confidentiality”

The above indicates that, to be determined as having taking corresponding confidentiality measures, the right holder is only required to match one of the listed cases provided it is sufficient under normal circumstances.

Article 13

If there is no substantial difference between the alleged infringing information and the trade secret, the court may determine that the alleged infringing information and the trade secret constitute substantially same as stated under Article 32, paragraph 2 of the Anti-Unfair Competition Law.

The court may consider the following factors when determining whether they are substantially same as mentioned in the preceding paragraph:

“…(3) Whether there are substantial differences between the use, usage, purpose and effect of the alleged infringing information and trade secrets”

This allows the right holder to reversely presume and judge the consistency at the source code level based on the preliminary comparison of the interface similarity, function and effect of the target program, and further combine the decompilation results of the target program, thus further reducing the burden of proof on the right holder’s side.

Suggestion: Appropriate measures should be taken from a technical point of view to have employees or other persons leave traces of ‘contact‘ on the company's trade secrets.

Persons in contact with confidential materials must be required to sign confidentiality agreements and non-competition agreements in advance. Such agreements should include specific and clear terms on the contents and carriers of trade secrets.

An operational software version control system could make it easier to identify the ownership of rights, and it is also helpful in that a most similar version can be elected for the comparison of ‘substantially same’ items.

It defines the ‘use’ of trade secrets.

Article 9

If the accused infringer directly uses trade secrets in production and business activities, or modifies or improves trade secrets, or adjusts, optimises or improves relevant production and business activities according to trade secrets, the court will determine that it belongs to the use of trade secrets as mentioned under Article 9 of the Anti-Unfair Competition Law.

This provision makes it clear that the degree of modification and improvement does not affect the nature of the infringement of ‘trade secrets’.

Suggestion: Protect software as a trade secret, so that the elements of behaviour of the alleged infringer are not as limited as in copyright infringement.


Jing Zhang

Patent attorney

Liu Shen & Associates

T: +86 10 6268 1616

E: Jing Zhang is a patent attorney at Liu Shen & Associates. Before joining Liu Shen & Associates, she worked as a software engineer at Telecommunication Research Institution of Samsung Electronics in South Korea for two years.

Jing specialises in patent application/inquiry, patent re-examination and invalidation, patent administrative and legal proceedings, and other matters related to IP rights.

Jing has a PhD from Shanghai Jiao Tong University.


Guang Hou

Patent attorney

Liu Shen & Associates

T: +86 10 6268 1616


Guang Hou is a patent attorney at Liu Shen & Associates. He specialises in patent prosecution, re-examination, patent invalidation, IP litigation, and client counselling with a focus on communication, computer networks, multimedia, electronics, and physics.

Guang has been providing legal opinions in litigation and consultation cases since 2011. In particular, he has worked independently in the fields of standard research, prior art searching, infringement analysis, and invalidation analysis in many cases. He is seasoned in FRAND licensing/infringement litigation and invalidation proceedings of standard-related patents.

Guang has a PhD from the University of Science and Technology of China (USTC). Before joining Liu Shen & Associates, he worked as a postdoctoral fellow at USTC for three years.

more from across site and ros bottom lb

More from across our site

Civil society and industry representatives met in Geneva on Thursday, September 28 to discuss a potential expansion of the TRIPS waiver
Sources say the beta version of the USPTO’s new trademark search tool is a big improvement over the current system but that it isn’t perfect
Canadian counsel weigh in on the IP office’s decision to raise trademark filing fees in 2024 and how they’re preparing clients
We provide a rundown of Managing IP’s news and analysis coverage from the week, and review what’s been happening elsewhere in IP
Shira Perlmutter, US Register of Copyrights, discussed the Copyright Office's role in forming generative AI policy during a House of Representatives hearing
The award marks one of the highest-ever damages received by a foreign company in a trademark infringement suit in China
Two orders denying public access to documents have reignited a debate over a lack of transparency at the new court
Rouse’s new chief of operations and the firm’s CEO tell Managing IP why they think private equity backing will help it conquer Europe
Brian Landry, partner at Saul Ewing, reveals how applicants can prosecute patent applications in the wake of the Federal Circuit's In re Cellect ruling
Ronelle Geldenhuys of Australia’s Foundry IP considers the implications complex computer technologies such as AI have on decision-making