Category Archives: European Patent Law

The EPO rejects two patent applications designating AI system DABUS as inventor

Posted on by
1

Abstract:

The EPO rejected two European patent applications, which both named an AI system as inventor, on the ground that the designation of inventorship did not fulfill the requirements of the EPC accepting only human persons as inventors.  The EPO states in the decision that AI systems or machines cannot be holders of the rights conferred to an inventor by the EPC.

A closer look at the two patent specifications reveals that these do not contain any evidence that AI was plausibly involved in the inventive process of devising the claimed inventions.

The recent decision of the European Patent Office (EPO) to reject two patent applications for failure to name a natural person as inventor has attracted attention recently. Earlier this week the reasons for the decisions to refuse EP 3 564 144 and EP 3 563 896 were published on the EPO register website.

The identical applicant of both patent applications claims that the true inventor of the inventions disclosed in the applications is an Artificial Intelligence (AI) system called DABUS, a machine intelligence system developed by the applicant based on artificial neural networks allegedly being able to perform creative work. The applicant designated “DABUS” as the sole inventor of both patent applications and stated that he had obtained the right to the European patent from DABUS as a successor in title.

The EPO refused both applications under Art. 90(5) of the European Patent Convention (EPC) for formal reasons, because the requirements of Art. 81 EPC

“The European patent application shall designate the inventor. If the applicant is not the inventor or is not the sole inventor, the designation shall contain a statement indicating the origin of the right to the European patent.” 

and Rule 19(1) EPC

“(1) The request for grant of a European patent shall contain the designation of the inventor. However, if the applicant is not the inventor or is not the sole inventor, the designation shall be filed in a separate document. The designation shall state the family name, given names and full address of the inventor, contain the statement referred to in Article 81 and bear the signature of the applicant or his representative.” 

were not met. These requirements were interpreted by the EPO in such a way that the designated inventor has to be a human person. Mentioning the name of a machine thus does not fulfill the requirements.

In defense of his request to designate the AI system DABUS as the inventor, the applicant relied in particular on the following arguments:

  • The Al system DABUS would be the actual deviser of the invention underlying the application. It would be a fundamental principle of patent law that the applicant must indicate the true inventor of the invention.
  • Rule 19(1) EPC would not require that the inventor is a human person.
  • Allowing an AI system or machine to be designated as inventor would be in line with the purpose of the patent system, namely to incentivise disclosure of information, commercialisation and development of inventions.
  • Not accepting Al systems as inventors would exclude inventions made by Al from patentability, contrary to Articles 52-57 EPC.

The EPO based its decision to reject the application essentially on the following reasons:

  • The legal framework of the EPC provides only for natural persons and legal persons to act within the system created by the EPC. Non-persons, i.e. neither natural nor legal persons do not have any role in proceedings before the EPO. In the context of inventorship reference is made only to natural persons. This indicates a clear legislative understanding that the inventor has to be a natural person.
  • Names given to things may not be equated with names of natural persons. Names given to natural persons serve not only the function of identifying them but enable them to exercise their rights and form part of their personality. Things have no rights which a name would allow them to exercise.
  • Under the EPC, the inventor has particular rights including the initial right to the European patent (which she can assign to a third party) and the right to be mentioned and designated as inventor on the publications of the application and the granted patent. Al systems or machines have at present no rights because they have no legal personality comparable to natural or legal persons. Legal personality is assigned to a natural person as a consequence of their being human, and to a legal person based on a legal fiction. Such legal fictions are either directly created by legislation, or developed through consistent jurisprudence. In the case of Al inventors, there is no legislation or jurisprudence establishing such legal fiction. It follows that Al systems or machines cannot have rights derived from being an inventor, such as the right to be designated as an inventor in the patent application. AI systems or machines also cannot transfer any rights such as the right to a European patent to a successor of title.

Consequently, the patent application was refused, because the designation of inventor filed by the applicant naming the machine “DABUS” as inventor does not meet the requirements of Art. 81 and Rule 19(1) EPC recited above. The decision is open to appeal by the applicant.

The EPO in accordance with Rule 19(2) EPC did not verify the claim that the AI system DABUS in fact is the inventor of the claimed invention. In the following we take a brief look at the two disclosures itself to check whether the AI system plausibly could be the inventor of the alleged inventions.

The first invention titled “Device and method for attracting enhanced attention” describes a device looking like a torch or pocket lamp having an LED light source 6 emitting light pulses having a fractal structure:

Claim 1 as amended after receipt of the search report from the EPO reads as follows:

  1. A device (2) for attracting enhanced attention, the device comprising:

(a) an input signal of a lacunar pulse train having characteristics of a pulse frequency of approximately four Hertz and a pulse-train fractal dimension of approximately one-half generated from a random walk over successive 300 millisecond intervals, each step being of equal magnitude and representative of a pulse train satisfying a fractal dimension equation of ln(number of intercepts of a neuron’s net input with a firing threshold)/ln(the total number of 300 ms intervals sampled); and

(b) at least one controllable light source (6) configured to be pulsatingly operated by said input signal;

wherein a neural flame is emitted from said at least one controllable light source as a result of said lacunar pulse train.

According to the description (paragraph [0017]), these particular light pulses defined in claim 1 should attract particular human attention:

“Even to the naked eye, and without the use of an anomaly detector, fractal dimension 1/2 pulse streams preferentially attract the attention of human test subjects. The most attention-grabbing aspect of such streams is that the ’holes’ or lacunarity between pulses occur as anomalies in what would otherwise be a linear stream of events. In other words, the pattern is frequently broken, such anomalous behavior possibly being detected by the TRN within the human brain as inconsistencies in the established arrival trend of visual stimuli. In contrast, should fractal dimension drop significantly below 1/2, the frequency of anomalous pulses drops, making them less noticeable to humans should either attention or gaze be wandering.“

The allegedly inventive light pulses having a fractal structure, however, are not generated by AI or machine learning but are simply received as an “input signal”. It is therefore not apparent at all, where the alleged machine intelligence system DABUS could have contributed to the invention. No evidence is submitted for the involvement of AI or neural networks or the like in conceiving the invention as defined by the above-recited claim 1. The claim features are not the result of a machine learning or other AI process. The allegation that DABUS is the real inventor seems not to be justified by the disclosure of the patent application.

The same applies to the second invention relating to a food or beverage container having a wall formed of fractal structures.

These fractal structures allegedly allowing coupling a plurality of such containers by inter-engagement, however, are likewise not described to be generated by AI. Again, a contribution of AI to the alleged invention is spurious.

The intention behind the two patent applications may actually be marketing-driven, namely to obtain a patent as an indication that the DABUS system indeed is able to perform creative tasks. The title “Device and method for attracting enhanced attention” then is aptly chosen.

EPO Enlarged Board of Appeal to Decide on the patentability of computer simulations

Posted on by
Reply

Summary

The following questions are referred to the Enlarged Board of Appeal for decision:

  1. In the assessment of inventive step, can the computer-implemented simulation of a technical system or process solve a technical problem by producing a technical effect which goes beyond the simulation’s implementation on a computer, if the computer-implemented simulation is claimed as such?
  2. If the answer to the first question is yes, what are the relevant criteria for assessing whether a computer-implemented simulation claimed as such solves a technical problem? In particular, is it a sufficient condition that the simulation is based, at least in part, on technical principles underlying the simulated system or process?
  3. What are the answers to the first and second questions if the computer-implemented simulation is claimed as part of a design process, in particular for verifying a design?

1. Introduction

The EPO case law on computer-implemented inventions has been quite stable for the last fifteen years. In particular, there has not been any decision of the Enlarged Board of Appeal (EBA) on this topic since G 3/08 of 12 May 2010.

Basically, a computer-implemented invention can be patented if it solves a technical problem by producing a technical effect. For the assessment of inventive step over the prior art only those features are taken into account, which contribute to the technical character of the invention.

There is, however, no general definition of what is “technical”.

The Technical Board of Appeal 3.5.07 (TBA) now has with decision of 22 February 2019 referred to the EBA the above three questions relating to the patentability of computer-implemented simulation methods. A decision by the EBA could have far-reaching consequences for applicants of computer-implemented inventions in Europe.

2. The invention

The invention disclosed in European patent application No. 03793825.5 and published as  WO 2004/023347 A2 relates to a computer-implemented method of modelling the pedestrian crowd movement in an environment such as a train station.

The modelling can be used to help design or modify the venue (e.g. train station) and for that purpose achieve a more accurate and realistic simulation of pedestrian crowds in real-world situations. The application is based on the insight that human interaction can be expressed and modelled in the same way as interactions of physical objects as e.g. electrons in a semiconductor device.

Claim 1 of the patent application according to the main request of the applicant reads as follows:

“A computer-implemented method of modelling pedestrian crowd movement in an environment, the method comprising:

simulating movement of a plurality of pedestrians through the environment, wherein simulating movement of each pedestrian comprises:

providing a provisional path (9) through a model of the environment from a current location (6) to an intended destination (7);

providing a profile for said pedestrian;

determining a preferred step (112′), to a preferred position (123′), towards said intended destination based upon said profile and said provisional path, wherein determining said preferred step comprises determining a dissatisfaction function expressing a cost of taking a step comprising a sum of an inconvenience function expressing a cost of deviating from a given direction and a frustration function expressing a cost of deviating from a given speed;

defining a neighbourhood (29) around said preferred position (123′);

identifying obstructions in said neighbourhood, said obstructions including other pedestrians (21) and fixed obstacles (25);

determining a personal space (24) around said pedestrian;

determining whether said preferred step (112′) is feasible by considering whether obstructions (21, 25) infringe said personal space over the course of the preferred step (112′).”

3. The Reasoning of the TBA

The TBA is of the view that the claimed invention lacks inventive step over a general purpose computer on the following grounds (section 11 of the decision):

“In the Board’s view, a technical effect requires, at a minimum, a direct link with physical reality, such as a change in or a measurement of a physical entity. Such a link is not present where, for example, the parabolic trajectory followed by a hypothetical object under the influence of gravity is calculated. Nor can the Board detect such a direct link in the process of calculating the trajectories of hypothetical pedestrians as they move through a modelled environment, which is what is claimed here. In fact, the environment being modelled may not exist and may never exist. And the simulation could be run to support purely theoretical scientific investigations, or it could be used to simulate the movement of pedestrians through the virtual world of a video game.”

In the proceedings before the TBA, the applicant has cited the earlier decision T 1227/05 relating to a computer-implemented method for the numerical simulation of an electronic circuit. T 1227/05 states: “An electronic circuit having input channels, noise input channels and output channels, the behavior of which is described by a system of differential equations is a sufficiently determined class of technical subjects, the simulation of which can be a functional technical feature.”

In other words, if the system or method to be simulated is sufficiently technical (expressed in engineering-heavy language ?), then the simulation method itself is also a technical method and for the assessment of inventive step has to be compared with prior art simulation methods, not with the functioning of a general purpose computer.

The TBA consequently held that if the board were “to follow decision T 1227/05, it would have to acknowledge that some or all of the steps of the simulation method of claim 1 contribute to a technical effect of the invention and could thus not be ignored when assessing inventive step” and that it “would hence be necessary to compare the invention with prior art other than a general-purpose computer.”

The TBA, however, still found that it “would tend to consider the subject matter of claim 1 of the main request to lack inventive step over a general-purpose computer” since the claimed method would “assist the engineer only in the cognitive process of verifying the design of the circuit or environment, i.e. of studying the behavior of the virtual circuit or environment designed.” And, “the cognitive process of theoretically verifying its design appears to be fundamentally non- technical.”

4. The referral questions

The TBA therefore referred the case to the EBA for reconciling its own view (citing a number of other earlier decisions as support) with the inconsistent view of T 1227/05. The questions are:

  1. In the assessment of inventive step, can the computer-implemented simulation of a technical system or process solve a technical problem by producing a technical effect which goes beyond the simulation’s implementation on a computer, if the computer-implemented simulation is claimed as such?
  2. If the answer to the first question is yes, what are the relevant criteria for assessing whether a computer-implemented simulation claimed as such solves a technical problem? In particular, is it a sufficient condition that the simulation is based, at least in part, on technical principles underlying the simulated system or process?
  3. What are the answers to the first and second questions if the computer-implemented simulation is claimed as part of a design process, in particular for verifying a design?

In effect, the question to decide here is if the required technical effect of a simulation method can be provided not only in the way the simulation program is running on the computer – then the relevant prior art is a general purpose computer – or if this technical effect can be provided also by technical properties of the simulated system (here: the design of the train station; in T 1227/05: the electronic circuit). If the design of a train station with staircases, exits, shops etc. is regarded in the same way as a technical subject matter as the design of an electronic circuit (which in my personal view should not be in doubt), then questions 1 and 2 have to be answered in the affirmative.

The outcome of the referral has in my view implications far beyond simulation programs, for example also to inventions based on machine learning (ML). If you replace in the referral questions “simulation” by “machine learning process”, the same issues arise.

There can be no doubt: If the first two questions of the referral are not answered in the affirmative by the EBA, applicants of various types of computer-implemented inventions in Europe will be in trouble.

 

Twitter: @patentlyGerman

IP5 offices CS&E pilot project: 500 PCT applications over the next 2 years will receive within 9 months an International Search Report jointly prepared by the five largest IP offices (CN, US, EP, JP, KR)

Posted on by
Reply

The world´s five biggest IP offices (collectively named IP5) started on July 1, 2018 a Collaborative Search and Examination (CS&E) pilot project to allow applicants of PCT applications upon request to receive an early search report including a provisional written opinion on patentability drawn up jointly by examiners of the Chinese, US, Japanese, Korean and European patent offices. This allows applicants to get an early assessment of the chances of success of a patent application based on a broader perspective of different prior art databases, search strategies, languages, and examiner expertise.

The applicant-driven procedure (see official illustration) starts with a request filed by the applicant with the Receiving Office, which will be transmitted to the International Bureau (IB) and the competent International Searching Authority (ISA), which has to be one of the IP5 offices.

IP5 pilot

The competent ISA will then assesses whether the requirements (see below) are met and will notify the applicant and the IB whether or not it accepts the request for participation in the pilot. The competent ISA (now main ISA) then carries out a search, prepares a search report and a provisional opinion on patentability. These are made available to the “peer” examiners at the other four IP offices through a safe ePCT-based platform. The peer examiners will provide the main examiner with their contributions, taking into consideration the provisional international search report and written opinion of the main ISA. The final international search report and written opinion will be established by the main examiner after consideration of the contributions from the peer examiners, which is in turn submitted to the applicant or its representative, if possible, within the time limit under Rule 42.1 PCT, i.e. within nine months from the earliest priority date of the PCT application.

Each of the participating offices will admit about 100 PCT applications to the pilot throughout a two-year period starting on July 1, 2018 such that in total 500 PCT applications will be participating. A common set of quality and operational standards will be applied by all collaborating offices. The CS&E pilot project is scheduled to continue until June 1, 2021 and will then be jointly assessed by the participating offices.

During the pilot phase the cost of the jointly prepared search report and written opinion is that of a normal PCT chapter I search at the competent ISA. The applicant thus gets the additional search and examination results from the other four offices for free. If the CS&E will be implemented as a regular product, however, higher fees are to be expected.

The requirements for participating at the CS&E pilot project are the following:

  • The request for participation in the pilot must be submitted using the standard form and filed together with the international application at the Receiving Office or the IB.
  • Until languages other than English are accepted into the pilot, the participation form and the international application must be filed in English.
  • One applicant cannot file more than 10 requests for participation in the pilot at a same main ISA.

If the pilot turns out to be successful, the collaborative work of peer examiners of the IP5 offices could bring a huge benefit for patent applicants. It may also allow a glimpse at a future international patent system harvesting international cooperation and standardization to avoid duplicated work and bringing faster, more cost-effective procedures for the applicants. Earlier knowledge about the patentability of an application also brings benefits to the industry as a whole and contributes to make the patenting process itself more transparent and efficient.

twitter: @patentlyGerman

 

CRISPR EP patent update: 7 oppositions filed against Berkeley CRISPR EP patent

Posted on by
Reply

The patent dispute relating to the revolutionary CRISPR-Cas9 gene-editing technology between the Broad Institute (jointly operated by Harvard and MIT) on the one side and UC Berkeley on the other side is is heating up in Europe.

On January 17, the Broad Institue lost its first granted CRISPR EP patent in first instance opposition hearings. An appeal against this decision has immediately been filed on January 18 by the patent proprietor.

On the other hand, the University of Berkeley (together with the University of Vienna and Emmanuelle Charpentier) filed it´s first patent relating to CRISPR technology even before the Broad Institute already on March 15, 2013 claiming the earliest priority on May 25, 2012. The resullting patent EP 2 800 811 was granted with main patent claim 1 as follows:

“1. A method of modifying a target DNA, the method comprising contacting the target DNA with a complex comprising:

(a) a Cas9 polypeptide and
(b) a single-molecule DNA-targeting RNA comprising:

(i) a DNA-targeting segment comprising a nucleotide sequence that is complementary to a sequence in the target DNA, and

(ii) a protein-binding segment that interacts with said Cas9 polypeptide, wherein
the protein-binding segment comprises two complementary stretches of nucleotides that
hybridize to form a double stranded RNA (dsRNA) duplex,

wherein said two complementary stretches of nucleotides are covalently linked
by intervening nucleotides,
wherein said contacting is in vitro or in a cell ex vivo; and
wherein said modifying is cleavage of the target DNA.”

A restriction to either eukaryotic or procaryotic cells (which played a major role in the parallel US interference proceedings) seems not to be present. The contacting of the target DNA to be modyfied by the method according to the claim, however, has to take place in vitro or in a cell ex vivo. 

The mention of grant of the patent was published on May 10, 2017.

Within the 9-month opposition term ending February 10, 2018 seven parties have filed an oppositions, namely:

Opponent 1: DF-MP Dörries Franck-Molnia & Pohlman
Opponent 2: Onno Griebling
Opponent 3: TL Brand & Co. Ltd.
Opponent 4: HGF Limited
Opponent 5: Jones Day
Opponent 6: Allergan Pharmaceutical International Ltd.
Opponent 7: Elkington & Fife LLP.

Of these opponents only one, Allergan Pharmaceutical International Ltd., is clearly identifiable, the other six oppositions are filed by IP firms acting as so-called “strawmen” for opponents which do not want to be identified. This practice explicitly allowed by the EPO is common in the pharmaceutical industry in Europe.

All opposition grounds are raised, namely lack of novelty and lack of inventive step, lack of enabling disclosure of the patented technical teaching, and extension beyond the content of the original application (added matter). Despite the huge amount of documents filed, a first instance decision in these opposition proceedings can be expected in approximately 18 months.

 

Disclaimer: Neither I myself nor the patent firm Betten & Resch, of which I am a partner, are representing clients involved in any patent cases relating to CRISPR-cas9 technology.

Twitter: @patentlyGerman

CRISPR EP Update: Broad Institute loses fundamental patent EP 2 771 468 at opposition hearing on 17 January 2018

Posted on by
2

The patent dispute relating to the revolutionary CRISPR-Cas9 gene-editing technology between the Broad Institute (jointly operated by Harvard and MIT) on the one side and UC Berkeley on the other side is followed by the biotech community with great interest .

Broad Institute has been the first one to get a patent issued on CRISPR-Cas9 technology in Europe, namely EP 2 771 468. Against this patent, 9 oppositions have been filed, most of which so-called strawman oppositions. The oral hearing finalizing the first instance of the opposition proceedings has been scheduled to take place on 16-19 January at the EPO in Munich. In the preliminary opinion issued together with the summons to the oral hearing, the oppisition division took a non-binding provisional view that the four priorities P1, P2, P5, and P11 cannot be validly claimed by the patentee, because the priority right had not been validly transferred to the applicants of PCT application forming the basis of the granted EP patent.

The decision was issued already in the early afternoon of the second day of the hearing: The opposition division maintained its position laid out in the provisional opinion and decided that the patent proprietor cannot validly claim priorities P1, P2, P5, and P11. As a consequence the claims as granted (main request) were revoked on the ground of lack of novelty. The 64 auxiliary requests filed by the patent proprietor were not admitted to the proceedings. The patent was thus revoked in full. The course of the oral proceedings can be seen in this twitter thread.

In short, the grounds for losing the essential priority rights are as follows: In order to validly claim priority right of a European patent application either directly filed or as a national phase of a PCT application (as in the present case), either (i) the applicant of the priority application and of the later application are identical or (ii) the applicant of the priority application assigns the application to the applicant of the later application, or (iii) the applicant of the priority application assigns the priority right to the applicant of the later application.

The assignments of the application or the priority right must have occurred before the filing of the later application (here: the PCT application), since, for reasons of legal certainty, a retroactive effect is excluded. However proof of such transfer may be provided later.

Furthermore, in case of more than one joint applicants of the priority application, each applicant of the priority application must either be also applicant of the later application or have assigned the application or the priority right to one of the applicants of the later application before the later application date. This was not the case for the priority applications P1, P2, P5, and P11. One of the applicants (and inventors) of these US provisional priority applications was no longer applicant of the PCT application and did not assign the right to the application or the priority right to one of the applicants of the PCT application. As a harsh consequence, the applicants of the PCT application and proprietors of the European Patent lost priorities P1, P2, P5, and P11. Without entitle-ment to these priorities, the subject-matter of the granted claims lacked novelty over the prior art on record.

The written decision of the opposition division can be expected in approximately two to three months. In a press release issued shortly after the decision on priority and before announcement of the full decision, the Broad Institute already announced to lodge an appeal against the finding of the opposition division.

Disclaimer: Neither I myself nor the patent firm Betten & Resch, of which I am a partner, are involved in any patent cases relating to CRISPR-cas9 technology.

(picture by sporst)   Twitter: @patentlyGerman

German Federal Court of Justice (FCJ) takes different view on undisclosed disclaimers than recent EPO Enlarged Board of Appeal decision G1-16 – FCJ “Phosphatidylcholin” (19 September 2017)

Posted on by
1

Unofficial Headnote:

Including a not originally disclosed negative feature (disclaimer) in a patent claim in order to delimit the claimed subject-matter from the prior art does not constitute an inadmissible extension of the subject-matter of the original patent application if the disclaimer does not have a technical effect.

To the relief of many patentees in particular in the chemical and biotech industries the Enlarged Board of Appeal has in its most recent decision G1-16 confirmed that undisclosed disclaimers are allowable under certain circumstances defined in the decision G1-03, namely in order to

  1. restore novelty by delimiting a claim against state of the art under Article 54(3) and (4) EPC (later published prior art);
  2. restore novelty by delimiting a claim against an accidental anticipation under Article 54(2) EPC; an anticipation being accidental if it is so unrelated to and remote from the claimed invention that the person skilled in the art would never have taken it into consideration when making the invention; and
  3. disclaim subject-matter which, under Articles 52 to 57 EPC, is excluded from patentability for nontechnical reasons.

An undisclosed disclaimer is defined by including into a patent claim a negative technical feature, whereby neither the disclaimer itself nor the subject-matter excluded by it have been disclosed in the application as filed (G1-16, point 14).

The Enlarged Board of Appeal (EBA) reasoned that such undisclosed disclaimers in most cases constitute added matter, i.e. “the requirements of Article 123(2) EPC leave virtually no chance of an undisclosed disclaimer being allowable” since “introducing any disclaimer per definitionem excludes subject-matter from a claim and, hence, changes the technical content of the claim” (G1-16, point 42). The EBO allowed undisclosed disclaimers under the above-mentioned clearly defined conditions despite the fact that they violate Art. 123(2) EPC, because under these circumstances a violation of Art. 123(2) EPC would not give the patentee or applicant an unwarranted advantage damaging to the legal security of third parties; this being the rationale behind Art. 123(2) EPC (G1-16, point 36).

The German Federal Court of Justice (FCJ), however, applies in the recently published ex-parte decision “Phosphatidylcholin” a different reasoning. In the case the FCJ had to decide upon an appeal on points of law (“Rechtsbeschwerde”) of the applicant of a patent application, which has been rejected both by the GPTO and the Federal Patent Court based on the ground that of an undisclosed disclaimer consisted of an inadmissible extension beyond the content of the original application. The patent claim before the FCJ reads as follows:

“Cosmetic, nontherapeutic use of

a) 5 – 30% by weight of sodium chloride and 

b) 5 – 30% by weight of glycerol,

in each case based on the total composition,

c) wherein these preparations are free from phosphatidylcholine

d) in hand protection creams, cleansing milk, sun protection lotions, nourishing creams, day creams or night creams for reinforcing the barrier function of the skin.”

The undisclosed disclaimer (feature c)) requires that the claimed preparations are free from phosphatidylcholine (a particular form of lecithin). Such preparations free from phosphatidylcholine are not disclosed in the original application documents. The GPTO and the Federal Patent Court therefore rejected the claim as inadmissibly amended, which finding was not contested by the applicant itself.

The FCJ, however, takes a different position and concludes that in the original application documents there is also no indication that phosphatidylcholine is a necessary component of the preparation or that its addition is considered advantageous. Therefore, the disclaimer “merely expresses the fact that from the large number of possible compositions of the preparation, which are possible according to the original application documents, only those containing phosphatidylcholine are excluded”. In addition, the undisclosed disclaimer merely serves to distinguish the invention over a (non-accidental, prepublished) prior art reference R disclosing skin protection preparations containing sodium chloride and glycerin in addition to the main ingredient phosphatidylcholine. The FCJ concludes that undisclosed disclaimers do not consist of an inadmissible amendment if (i) they serve the purpose to distinguish the claimed subject-matter over the prior art and (ii) the disclaimer has no technical effect. The use of undisclosed disclaimers is allowable during the examination as well as in post grant procedures.

The reasoning of the FCJ thus differs from that of the EBA. The EBO concludes: Undisclosed disclaimers constitute an inadmissible amendment but are exceptionally allowable under the above-mentioned circumstances 1., 2., or 3. The FCJ finds: Undisclosed disclaimers do not constitute an inadmissible amendment in the first place and are therefore allowable (i) if they serve to distinguish the claim over the prior art and (ii) if they do not have a technical effect.

In practice, the main difference is that under the FCJ approach the allowability of undisclosed disclaimers is also applicable to non-accidental prepublished prior art (as reference R in the subject decision “phosphatidylcholine”). Condition (ii) requires that the undisclosed disclaimer must not have a technical effect, i.e. the skilled person must not derive any new technical information therefrom. This condition seems to imply that undisclosed disclaimers are not be accepted to support inventive step. The FCJ has decided in the “phosphatidylcholine” case that the disclaimer does not have any technical effect based on the application documents without reference to the prior art, such that the question of inadmissible extension is independent from the considered prior art.

The assessment of a potential technical effect of the undisclosed disclaimer, however, may lead to the following problem:

In the present appeal on points of law case the FCJ had to refer the matter back to the Federal Patent Court to decide on the merits whether the claim containing the disclaimer is novel and involves an inventive step over the prior art. It is to be expected that novelty will be confirmed, because otherwise the disclaimer would not have made any sense. More interesting, however, is the assessment of inventive step: The closest prior art document R discloses a skin protection preparation comprising phosphatidylcholin (as main, essential ingredient) + glycerol + sodium chloride (features [C+A+B]). The application under review claims glycerol + sodium chloride + (not phosphatidylcholin), i.e. features [A+B+notC]. Would it be obvious for the skilled person starting from closest reference R to provide a skin protection preparation removing the main, essential component? Probably not. In this case, however, the disclaimer notC would indeed have a technical effect establishing inventive step, contrary to the finding of the FCJ allowing the disclaimer in the first place. It will be interesting to see how the Federal Patent Court will resolve this matter.

The EBA approach, in contrast, appears to avoid such problems: A negatively claimed technical feature is regarded as (almost) always having a technical effect such that its addition contravenes Art. 123(2) EPC, but is exceptionally allowed under certain precisely defined conditions avoiding unnecessary hardship for the patent proprietor while preserving legal certainty for third parties.