A blog series by Zoltán Gyenge and Thomas Wyder
(Software Practice Group at Rentsch Partner)
with emphasis on the implications of the recent landmark decision G 1/23 of the Enlarged Board of Appeal of the European Patent Office for Inventions in the fields of Information and Communication Technology (IT, computer implemented inventions, software, quantum computing, artificial intelligence, …)
Part I – “Grey´s Anatomy of Prior Art”
The Black and White Era (1992-2025)
So far, the general understanding was that there is one standard for “disclosure” under the European Patent Convention (EPC), i.e., one “disclosure test”.
After all, the EPO´s Guidelines, G-IV.2 state: “Subject-matter can only be regarded as (…) comprised in the state of the art pursuant to Art. 54(1), if the information given is sufficient to enable the skilled person, at the relevant date (see G‑VI, 3) and taking into account the common general knowledge in the field at that time, to practise the technical teaching which is the subject of the disclosure (…).”
Possibly the most important legal foundation in this context was set by the landmark decision G 1/92, which established that a product becomes part of the state of the art when it has been made publicly available and can be analyzed and reproduced. The headnote of G 1/92 states:
- “The chemical composition of a product is state of the art when the product as such is available to the public and can be analysed and reproduced by the skilled person, irrespective of whether or not particular reasons can be identified for analysing the composition.
- The same principle applies mutatis mutandis to any other product.”
In this view, both availability and reproducibility were required. G 1/92 was often read as excluding from prior art any product that could not be analyzed to the point of reproducibility
It thus appeared safe to assume that either a technical teaching constitutes prior art, as it was fully disclosed, i.e., disclosed in an enabling manner for the skilled person to put an invention into practice, or there is, from the skilled person´s point of view, ´nothing to see` (no prior art) at all: a black and white approach towards prior art.
This black and white approach was commonly accepted to relate to all types of disclosure and, hence, to all types of prior art: written and oral publication, as well as public prior use in terms of products put on the market.
The “disclosure test” for prior art and for enablement appeared to be one and the same. However, this was not the end of the story.
The Era of Shades of Grey (2025-)
On July 2, 2025, the Enlarged Board issues a decision, G 1/23, introducing a significant refinement to the concept of disclosure in European patent law, which can be understood to introduce shades of grey to disclosure.
The Enlarged Board clarifies that the headnote of G 1/92 addressed only the situation where reproducibility is possible. It did not imply that lack of reproducibility automatically excludes a product from the state of the art. Instead, G 1/23 affirmed the following:
- “ A product put on the market before the date of filing of a European patent application cannot be excluded from the state of the art within the meaning of Article 54(2) EPC for the sole reason that its composition or internal structure could not be analysed and reproduced by the skilled person before that date.
- Technical information about such a product which was made available to the public before the filing date forms part of the state of the art within the meaning of Article 54(2) EPC, irrespective of whether the skilled person could analyse and reproduce the product and its composition or internal structure before that date.”
In G1/23, one of the guiding motives for the Enlarged Board´s conclusions lay in the ambition to avoid legal fictions whenever not explicitly required by the law (here: the EPC): “(…) the legally not existing but otherwise commercially available product that is even deliberately brought to the attention of skilled persons can only be regarded as a legal fiction. Legal fictions that override facts do exist in law, including patent law, but they should normally be explicitly stated in the law. These considerations alone dictate that the fiction of the exclusion of the product from the state of the art should be treated with serious reservations.” (G 1/23, Reasons 48).
Without explicitly saying so, it is fair to conclude that the Enlarged Board in G1/23 distinguishes between two types of disclosures associated with products put on the market:
- “enabling disclosures” that allow the skilled person to reproduce the invention, and
- “de facto disclosures” of real-world products that may not be fully analyzable and reproducible (e.g., its internal workings cannot be fully reverse-engineered), yet still form part of the state of the art, as they can (or could) be physically obtained.
Thus, one can conclude that the Enlarged Board introduces a double standard for the disclosure of factual products: a (lower) standard of merely being factually ´put out there` and thus becoming prior art and the (higher) standard for reaching an enabling disclosure, the higher standard being a requirement to be met when seeking a patent.
G 1/23, therefore, introduces a unique treatment for real-world ("factual") products, separating them from more traditional forms of prior art, i.e., written or oral disclosures.
The Limits of Analyzability and The Resulting Grey´s Period for Filing
The concept of partial analyzability of a product plays a central role in G 1/23. A product may have aspects that are transparent to the skilled person and aspects that remain opaque: a partial black box, i.e., a grey box. Even in the extreme case of a complete black box (e.g., a secret formula like Coca-Cola, an example mentioned by the Enlarged Board), the product itself is, against the backdrop of G 1/23, to be considered prior art.
What constitutes state of the art, at any given point in time, are only those features of the product which would have been analyzable using common general knowledge (CGK) available at the relevant date, irrespective whether said analysis ever took place or not. Later advances in common general knowledge with respect to analysis tools may bring the potential of revealing more features as prior art. However, to be dogmatically consistent, such further features must only enter the prior art ex nunc, but not ex tunc, i.e., not retroactively. Thus, when further features of the product are uncovered using the new tools of analysis entering common general knowledge, it appears that a further piece of prior art comes into being, which is relevant for any patent application filed after the new analysis tools became available: The product with the (now newly) analyzable (larger) set of features. Thus, the shade of a grey box changes.
Products can be de facto disclosed in a non-enabling manner and constitute prior art (albeit less significant prior art than if they would have been disclosed in an enabling manner), but there is a higher standard to be met for an enabling disclosure. Moreover, when more information becomes known about a certain product, whether through active disclosure or because an advance in common general knowledge allows to obtain more information through analysis, further shades of grey appear and become relevant at a certain date. Put differently, a product can repeatedly become prior art with increasing levels of detail being included, i.e., public prior use considered at different points in time is, hence, associated with different shades of grey.
Eventually, when a disclosure becomes fully enabling, the grey box turns into something transparent, thus fully exposing the interior life of the product. For as long as a disclosure remains non-enabling, as the product is not fully analyzable and reproducible, thus remaining a grey box, this situation gives rise to a grey´s period (sic!) during which competitors can still file an application directed at the hidden aspects of the product.
The Chain of Prior Products Leads Back to Nature
The heart of the Enlarged Board´s reasoning in G 1/23 appears to lie in a philosophical approach. Instead of insisting on legal evidence for reproducibility and treating the question whether a product has been disclosed as a legal question, the Board adopts a naturalistic perspective and exclusively asks the factual question of what has been made available to the public, irrespective of whether it is sufficiently complete to constitute a technical teaching.
In a nutshell, the Enlarged Board reasons that every product depends on some starting material and said starting material depends on its own starting material, i.e., there is a chain of dependency of starting materials leading back to the primary starting material, said primary starting material, according to the Enlarged Board, lying in nature, namely, in atoms. After all, according to the Enlarged Board, the world offers products (like crude oil or minerals) that exist independently of human intervention, and these products can be used, even if they cannot be synthesized. The intriguing detail here is that the Enlarged Board does not distinguish between products offered by nature (e.g., crude oil, minerals) and man-made products (such as Coca-Cola), even although the EPC appears to hint at a distinction by carving out so-called discoveries from the realm of possible inventions.
Since the primary starting materials of nature are non-reproducible (the Enlarged Board mentions that humans cannot create atoms; Reasons 60), insisting on reproducibility as a precondition would, according to the Enlarged Board, eliminate all prior art. The Enlarged Board concludes that requiring full reproducibility of all components of a product would result in the prior art being the mathematically empty set, a reductio ad absurdum (Reasons 65; see also Reasons 67, 70). Consequently, the Enlarged Board deduces that reproducibility is not a necessary criterion for a product to be part of the state of the art (see also Reasons 73).
The Impact of Grey Boxes on Patenting ICT
There is reason to note that G 1/23 reshapes several aspects of patent practice. Inventions in the field of ICT may feel the impact of grey box products as prior art most acutely. Where software is embodied in a product, but its internal logic is not reproducible, G 1/23 may limit the room for subsequent protection. In contrast, new considerations arise when deciding to keep the internal life of a product a trade secret. Selling a product without disclosing its internal structure in a reproducible manner still blocks others from patenting the product itself later on. However, it is to be kept in mind that the grey box product would not prevent a competitor from inventing and later on protecting a (even the same) previously undisclosed method of production.
Many products in the field of ICT can be put on the market without laying open their internal structure and functioning to the extent that these could be reproduced. This applies to digital products, such as chips, embedded systems, and AI models, exposing observable technical features while concealing the means of their creation. These products are often analyzable using tools and methods available at the relevant date, allowing the skilled person to derive certain behaviors or structures, even though the underlying design, training process, or manufacturing method may remain undisclosed or irreproducible. Under G 1/23, such partial disclosures (grey boxes) can still destroy novelty, even if reverse engineering the product is unfeasible or the internal logic remains a black box. This asymmetry hits ICT inventions particularly hard: they are frequently commercialized in non-transparent forms, while still revealing just enough to fall within the prior art, creating a broader field of novelty-destroying disclosures without relaxing the strict enablement requirements for patentability.
Coming up Next
In upcoming blog posts, we will explore the impact of G 1/23 on ICT inventions, including in particular digital and software-related inventions. The posts will analyze concrete scenarios where the decision’s distinction between factual availability and enablement reshapes how prior art is assessed under the EPC.
In particular, we will explore topics such as:
- What if a marketed device performs novel functions, but its firmware is locked away in read-only memory?
- Chips that can be observed but not manufactured: Does physical analyzability without fabrication know-how count as disclosure?
- When system behavior is public but backend logic is hidden (e.g., in a cloud-based system), is anything really disclosed?
- Is a downloadable grey box neural network / a trained model prior art?
- A proprietary quantum chip or algorithm is commercially available, but its internal workings, architecture, qubit control mechanisms, and/or error correction code, are not reproducible: What can a later inventor of the same quantum chip still patent?
Conclusion
G 1/23 marks a pivotal development in European patent law. While it does not overturn G 1/92, it carves out a nuanced category of de facto, non-enabling disclosures for real-world products. In doing so, it acknowledges the complex interplay between factual availability, technological limitations, and the evolving understanding of what constitutes the state of the art.
By elevating what exists over what can be reproduced, G 1/23 opens a new chapter in how we think about knowledge, disclosure, and innovation in the patent system. The message which has been broadcast by G 1/23 is clear: even a black or grey box can cast a long shadow on patentability.
