Trade Secrets As Supplement To Patent Protection¹

While complementary trade secret protection offers competitive advantages that extend far beyond the limited term of a patent, it complicates licensing as licensees lack access to crucial know-how that is required to effectively implement the invention.

I. Introduction

Traditionally, innovators had to choose between filing a patent application or keeping their invention secret. Nowadays, many cutting-edge technologies, particularly in the field of artificial intelligence (AI), rely on both protection mechanisms at the same time. AI algorithms can often be patented,² whereas training datasets are kept as trade secrets. This hybrid protection model presents opportunities and challenges. While complementary trade secret protection offers competitive advantages that extend far beyond the limited term of a patent, it complicates licensing as licensees lack access to crucial know-how that is required to effectively implement the invention.

Secrecy can slow down innovation and limit the monetization of protected technology and information. To address this issue, we propose the creation of a Trusted Depository (TD). This would reduce transactional risks, promote collaboration, and enable the commercialization of secret know-how while increasing legal certainty. This article examines cases in which trade secrets can complement patent protection (II.), explore the downsides of additional trade secret protection (III.), and lay out the idea of a TD (IV.).

II. Cases in which Trade Secrets can complement Patent Protection

1. Training Data

AI systems require vast amounts of training data, which are expensive to collect and organize.³ Besides legal hurdles that make it difficult and costly to collect certain types of data in the first place, transforming raw data into usable training data can involve quite a bit of “manual labor”: Data must be checked for quality.⁴ Missing data points must be tracked down,⁵ estimated and added.⁶ Faulty data points need to be corrected.⁷ Finally, the dataset must be tailored to be representative of reality. Even when the teaching of the patent itself has become public knowledge, it can remain prohibitively expensive to acquire the training data required to replicate the invention’s effects. This greatly reduces the attractiveness of a patent license. Ideally, training data should be easily tradable as an “add-on” to patent licenses. Besides the obvious benefits to patent holders and licensees, wider implementation of a given technology significantly increases the likelihood of improvements and follow-on innovations.

2. Additional Know-How

Valuable datasets and additional know-how are often also acquired by inventors of other technologies. While patents disclose solutions to technical problems, these solutions are rarely perfect and can often be improved with more practical experience. Slight changes in material, temperature, angle etc. can result in significant efficiency gains that can make or break the economic viability of a given technology. Additional know-how is not always patentable and can often only be protected through secrecy. The licensing of such technologies becomes difficult and risky, even though their widespread adoption would be in everyone’s best interest.

III. Downsides of supplemental Trade Secret Protection

Currently, secrecy heavily incentivizes companies to adopt a “closed innovation” strategy, prioritizing internal, non-cooperative R&D activities without sharing the results with other industry participants. This has pro-competitive aspects, as it inevitably leads to a greater divergence in available technology, but they rarely outweigh the downsides. In a closed innovation environment, a significant share of overall investment goes towards developing solutions to problems that others have already – secretly – solved. Duplicate inventions carry a significant societal cost. If the inventor had been aware of the existing solution and had been able to license it, they might have directed their resources toward improving it or solving a new problem entirely. This also applies to AI training data, which, instead of being gathered and compiled by numerous companies in parallel, should be shared wherever possible.

IV. Trusted Depository

The establishment of a TD could significantly improve the current situation by providing the necessary infrastructure and safeguards for licensing. Trade secrets are inherently difficult to license. Currently, potential parties to mutually beneficial licensing agreements are very unlikely to become aware of each other’s existence. A TD could include a browsable database that catalogues and describes the nature of deposited datasets and secrets. Depositors could decide freely whether they want to be listed in the database, which would also indicate if a deposited piece of information is available for licensing and who to contact to enter negotiations. This would stimulate licensing activity, shifting the market towards an “open innovation” approach.

A TD could also serve evidentiary purposes, especially when trade secrets are used to protect otherwise patentable inventions. While secrecy has a legally unlimited term of protection, anyone who discovers the secret through legitimate means can file a patent application and, since undisclosed secrets cannot constitute prior art, would likely be granted a patent. In most jurisdictions, the trade secret holder can claim a right of prior use, if he can prove that he had access to the invention before the priority date of the patent and had at least made the necessary preparations to begin using it.⁸ Creating and depositing a detailed account of his secret technology allows the depositor to reliably prove his earlier possession of the technology and the necessary intent and preparations to implement it. Such deposits also allow secret holders to file patent applications in a very short time. If a deposited secret has been compromised, the ability to quickly file a patent application and claim priority for the invention can be a valuable back-up plan.

Blockchain technology should play an important role in establishing a TD. By combining encryption, hashing and decentralized verification mechanisms, blockchain technology can guarantee confidentiality while providing a legally certain way to prove the authenticity and integrity of previously deposited information. To ensure secure and seamless execution of licensing agreements, smart contracts (self-executing agreements) could be integrated into the TD.⁹

The basic technological infrastructure for a TD is likely to be in place very soon anyway. The EU AI Act requires certain AI providers to register their AI in an EU database and undergo conformity assessments,¹⁰ which will generally require disclosure of all training data. Similar to the proposed TD, the database will have a publicly searchable section and a secure, non-public section accessible only to relevant authorities. Permitting the deposition of other contents, especially trade secrets, would be an important step towards establishing a TD.

V. Summary and Outlook

While patents provide a well-defined mechanism for safeguarding innovations, additional know-how often determines the economic viability of an invention. Such know-how is typically kept secret. This complicates licensing and commercialization, hindering the widespread adoption and further development of protected technologies. Despite clear economic incentives, trade secret licensing remains underdeveloped due to legal uncertainties and practical difficulties in maintaining secrecy while enabling controlled access. The proposed TD helps address these challenges. In addition to serving as verifiable proof of existence and ownership, it would enable potential licensors and licensees to connect via a searchable database. The ongoing development of regulatory frameworks, especially the EU AI Act, presents an opportunity to align the TD concept with emerging legal standards. The AI Act’s envisioned database could provide the required technological infrastructure and serve as a foundation. The successful implementation of a TD will require close collaboration between policymakers, industry leaders, and technology experts to ensure legal robustness and practical usability. While establishing such an infrastructure presents technical and legal challenges, many of the necessary building blocks — including secure digital storage and blockchain verification — are already available.

Disclaimer – The views expressed in this article are the personal views of the authors and are purely informative in nature.