"Recap and Update:

The first half of this article, [Guest post] Has G1/23 harmoniously "recast" the European law of enablement? noted that the Enlarged Board of Appeal (‘EBA’) in G1/23 was observed to have reinterpreted G1/92 so that reproducibility is fulfilled by market availability of the product for the product to become part of the state of the art. The Board also decided not to consider what an ‘undue burden of analysis for a person skilled in the art’ would be. 

On 2nd October 2025, the referring Board provisionally affirmed and applied G1/23 in its provisional opinion T438/19, reason 8.2. The Board considered that the prior publicly available product is part of the state of the art, reason 8.3, and is available as prior art for assessment of the patentability of the claimed invention, reason 10.5. The Board had the view for this specific case that a person skilled in the art would be able to prepare another material having similar properties of the irreproducible commercially available product without the need to find out the exact way of making it, reason 8.7.3, 4th paragraph. At the oral proceedings of 15 October, the respondent withdrew its consent and agreement to the text of the patent as granted and withdrew its pending requests. The final decision, T438/19, is therefore silent on the application of G1/23 by the board; the provisional opinion is at best commentary.

However, in T 0807/23, G1/23 has been applied: the Board of that case decided that the case should be remitted to the Opposition Division for first instance to consider the use of non-reproducible prior art in relation to inventive step, as provided for in G1/23, since such considerations were not available to the parties and the opposition division at the time of the decision under appeal.  Further remission to the Opposition Division would permit consideration of whether a technical data sheet for the commercially available product is part of the state of the art (Art 54(2) EPC).

It appears G1/23 has rendered established case law coherent.  However, is this applicable to all technical fields and has this decision been made in view of the downstream courts, G1/24, reason 16?

Other Technologies

A referral to the EBA is for the purposes of ensuring uniform application of the law (Art 112(1) EPC) and should, in theory, apply to all technical fields. Yet the underlying case and the applied case law appears to be limited to the ‘structure and composition’ of material products. In many reports of G1/23, its implications are only considered in this and related technical fields of chemistry and materials. Are these the only technical fields which are impacted by the decision, or does the decision have relevance for other technical fields such as software as implied by Art 112 EPC?

As noted, G1/23 has no broadening statement as G1/92. Since G1/92 is applicable to all technical fields, using G1/23 to recast G1/92, it can be argued that G1/23 thus indirectly applies to all technical fields.

Yet that might be too simplistic.

In the proceedings of G1/23, Forresters’ Amicus Brief (see page 1) of 8 December 2023 (the ‘Amicus Brief’) notes that the referral question addresses the chemical technical field and all the cited case law is from the same technical field, as noted by the Opponents (paragraph 103, Opponent’s letter of 16 October 2024). The Amicus Brief notes that the answers were likely to have a far wider impact in many technical fields such as computer implemented inventions, which is not explicitly referred to in G1/92. (G1/92 explicitly mentions mechanical or electrical articles, page 2, first paragraph of the Amicus Brief).

The Amicus Brief notes and reviews the case law on enablement for computer implemented inventions, culminating in T2440/12. This decision considers the limitations of decompilation and consequently analysability of software in Europe in view of copyright law. Decompilation of a software product in Europe is limited to a number of exceptions. Unconditional decompiling or disassembling of computer programs is not permitted. A skilled person would be deemed to be infringing copyright protection in software when decompiling it to gain knowledge of the underlying algorithm. Despite this limitation being inapplicable internationally, beyond Europe, the Deciding Board accepted this limitation (reason 12, T2440/12).

Yet observing, studying or testing the functioning of a program in order to determine the ideas and principles which underlie any element of the program without authorisation of its copyright owner is a permitted exception (Article 5, paragraph 3 Directive 91/250/EEC, replaced by Directive 2009/24/EC) (reason 12.3, T2440/12). Such analysis may be done while performing any of the acts of loading, displaying, running, transmitting or storing the program, i.e., without explicit decompilation and infringement of the copyright. The Deciding Board accepted that it is known in the art that such analysis without decompilation can be achieved line-by-line on a virtual machine.

This approach would also reveal the method steps of the software (reason 12.4, T2440/12). By executing the instructions line-by-line, the skilled person can derive knowledge of all the operations to be performed in order to carry out the method embodied by the software product (reason 13.2).  Thus a program's structure and functionalities could be revealed to the skilled person in a very concrete manner without infringing copyright in the software. The decryption prohibition is not, as such, a limitation to the analysis a person skilled in the art may make.

So long as the functionality and/or structure of the prior publicly available software is the same as that claimed, the patent claim will lack novelty (reason 13.4), even if the implementation in the product is different from that described in the patent.

In making this observation, the Deciding Board presents similar exemplary analysis between a patent claim to an electrical circuit and the disclosure of a real circuit placed on the market before the filing of the patent (reason 13.1). The real circuit would be prior art by virtue of its analysability by the person skilled in the art and anticipate the patent claim if a person skilled in the art found it (and its components) had the same functionality, even if the real circuit was realised differently from a circuit disclosed in the patent description (reason 13.3).

A bottom-up analysis thus appears to confirm that G1/23 is consistent with the practice in software and other technical fields. Considering the practice of other technical fields, G1/23 appears to make the approach to enablement in the chemical technical field consistent with the existing practice of other technical fields.

Harmonisation of Prior Product Disclosure before the Courts?

The EPC and its interpretation is a source of substantive patent law of the UPC as well as the EPO, see UPCA Art 24(1)(c). The implications of this were recognised in the recent EBA decision G1/24, the seminal decision on the use of the description on claim interpretation, which advocated greater harmonisation between the practice of the European Patent Office (the EPO) in interpreting the EPC and its downstream courts (reason 16, G1/24).  Has G1/23 sought a further step towards harmonisation long sought for by the fathers of the UPC and the President of the Boards of Appeal, although the decision G1/24 was decided just a week before that of G1/23, and despite the decision neither citing G1/24 nor case law of the national courts or of the UPC?  

The absence in the decision of a reference to court decisions might seem curious considering prior use disclosures are more likely to be considered by the courts as prior art than the proceedings at the EPO where documentary prior art is normally considered.   

A national court decision was considered in the referral decision. Reason 16, T0438/19 referred specifically to Takeda UK LTD v F. Hoffmann-La Roche AG [2019] EWHC 1911 (‘Takeda’) which relates to a biotech invention. The Referring Board observed the English court in Takeda identified the same divergent EPO case law as the EBA in applying G1/92, including the ‘full reproducibility’ branch. The judge, Birss LJ, favoured a different divergent branch: the contribution to the state of the art of an older product on the market is ‘the information the skilled person can obtain by analysing the older product is enough to enable the skilled person to put it into practical use by making their own version’ (para 130, Takeda).

What is curious is that Birss LJ asks himself in passing (i.e., obiter) if the law needs the product to be reproducible:

… since if the product really is freely available in as large a quantity as anyone would want and no-one actually needs to make it themselves, one might say it had been "made available to the public" (Art 54(2)) para 121, Takeda

His observation appears so close to the decision of G1/23 as almost to anticipate it, but for the fact it is not part of the reasoning (i.e., ratio) of his decision.

Takeda is the only court decision referred to in the proceedings. If it does represent the position of the national courts and the UPC on this question of the law, it would appear that EBA has taken note of the practice before the downstream courts and sought harmonisation.

And Harmonisation for Inventive Step?

There is one small wrinkle: the commercially available product in Takeda is considered for novelty, yet at para 135 Birss LJ notes such a prior art could exist for inventive step. Since the use of the disclosure of the prior art product is referred to for inventive step in G1/23, it would appear the EBA answered Birss LJ’s question, and the referring Board is likely to apply and affirm it, see the Board’s preliminary opinion of T438/19 (reasons 8.3 and 10.5).

Conclusion

As noted in the summary for the core question, although the order of G1/23 appears to be clear in itself, another approach to interpreting decisions of the Enlarged Board as applied in G1/92 in which the reasoning of the decision is read with the order, the impact is more nuanced.

G1/92 has been rewritten by G1/23 so the reproducibility requirement is silent – it may be fulfilled by the availability of a product on the market.

The recast test is a question of analysability unless unduly burdensome to a skilled person. What would be ‘unduly burdensome’ for the person skilled in the art to analyse has not been answered because the EBA decided that answering that question neither answered any referral question nor assisted the Referring Board addressing the case from which the referral derived. This remains an open question perhaps for a future referral.  

Without an answer to this question, G1/23 suggests that a product placed on the market is analysable without limitation: what a skilled person may derive from analysis of a product on the market is disclosed for the purposes of patentability. This is not a hard-and-fast on-sale bar, but a precautionary interpretation subject to developments in the law. Until the law is clearer, a risk remains that any feature disclosed to a skilled person by their analysis of the product could be part of the state of the art. So, for now, it is advisable for a proprietor of a product who may wish to file patent applications about the product and its later improvements, to take precautions in placing their product on the market and in releasing marketing materials about the product.

Although the referral questions of G1/23 directly related to the chemical technical field, review of the case law for software suggests that the decision appears to have brought practice in such diverse technical fields into coherence. In any event, by recasting G1/92 – which is applicable to all technical fields - G1/23 as such has implications for all technical fields.

G1/23 is not the end of the road but appears as a step along the path of harmonisation. Despite an absence of explicit citations from UPC and national court decisions, G1/23 appears to have been made with the downstream courts in mind. How the Boards of Appeal continue along this path with the courts will be keenly watched."

Source: https://ipkitten.blogspot.com/2025/11/guest-post-has-g123-harmoniously-recast.html