Artificial intelligence is not breaking patent law: EPO publishes DABUS decision (J 8/20)

The EPO Board of Appeal has published its full decision on the question of whether a machine can be an inventor (J 8/20). The Board of Appeal had previously announced its decision to refuse two European patent applications naming an algorithm ("DABUS") as the sole inventor at the end of last year (IPKat).

The decision in J 8/20 demonstrates that the current patent system is more than capable of dealing with AI inventions when and if they arise, without harming innovation or treating the AI inventors unfairly. Thus, contrary to the recent Nature article on this topic, AI is not breaking patent law. 

Case Background

J 8/20 relates to appeals of the decisions to refuse two European patent applications (EP3564144 and EP3563896) for failing to designate a natural person as an inventor. The patent applications purportedly relate to the inventions of an AI ("DABUS"). Whilst the AI was specified as the inventor, the patent applicant (and self identified owner of the inventions) was Dr Thaler. 

Succession of title

One of the fundamental issues with AI inventors is establishing the chain of title from the invention to the applicant for the invention from an inventor (the AI) lacking legal personality. Whilst DABUS is said to be capable of inventing, it does not yet seem capable of applying and owning a patent for its invention. This responsibility has been taken on by the inventor of DABUS, Dr Thaler. Dr Thaler argued that he derived the right to DABUS's inventions by virtue of being the AI's employer [Merpel: Raising the question of whether DABUS gets decent pay and holiday? Has anyone asked DABUS if it wants its inventions to be patented?]. 

The EPO Receiving Section rejected the applications on the ground that the applicant's right to the invention had not been established (IPKat). Particularly, DABUS lacks legal personality and so cannot transfer rights to the inventions to Dr Thaler (Article 60(1) EPC and Article 81 EPC). The Board of Appeal agreed with this assessment, noting that "under the EPC the designated inventor has to be a person with legal capacity. This is not merely an assumption on which the EPC was drafted. It is the ordinary meaning of the term inventor". 

Innovation and fairness

A central tenet of the AI inventor team's argument is that denying applicants the right to name an AI as an inventor will stifle innovation. However, this argument ignores the simple workaround that, under European patent law, the user or the owner of a device involved in an inventive activity can just designate themselves as inventor (r. 4.6.6). 

The Board of Appeal in the case in question also rejected the applicants argument that the public has a right to know who the inventor is and how the invention was made. The Board of Appeal noted that an inventor has the right for their name to be withheld from publication, and that a third party has no rights in this respect (r. 4.3.6). 

The Board of Appeal concluded that they were convinced that there is not a  problem of unequal treatment for computer assisted inventions calling for an evolutive interpretation of European patent law. The Board of Appeal particularly was not aware of any case law which would prevent the user or the owner of a device involved in an inventive activity to designate himself as the inventor under European patent law. 

If an applicant wished to acknowledge the AI inventor out of fairness, they are free to do this in the description (r. 4.3.7). If DABUS is an inventor, the real question of fairness is therefore how Dr Thaler is planning to compensate DABUS in exchange for Dr Thaler's exploitation of DABUS's inventions absent an employment contract between Dr Thaler and DABUS, that DABUS freely entered into. 

Final thoughts

"Salvation"

While philosophical discussions of AI inventorship may soon become pertinent as the field develops, the Board of Appeal in J 8/20 found no legal, moral or practical difficulties with dealing with such inventions under the current system. 

As a final observation, it is a shame that the present debate over AI inventorship is based on questionable science (IPKat). Dr Thaler claims that his algorithms are not just capable of inventing, but that they are also capable of sentience and near-death experiences. As long ago as 1996, Dr Thaler also claimed that his algorithms provide credible evidence for a "virtual afterlife", and the religious concepts of salvation and damnation. The Board in J 8/20 stressed that the appeal did not relate to whether an AI, or DABUS in particular, is capable of inventing. Unlike the courts and patent offices, however, the willingness of the patent community to take Dr Thaler and his "imagination engines" seriously is rather embarrassing.

Further reading

Artificial intelligence is not breaking patent law: EPO publishes DABUS decision (J 8/20) Artificial intelligence is not breaking patent law: EPO publishes DABUS decision (J 8/20) Reviewed by Rose Hughes on Monday, July 11, 2022 Rating: 5

7 comments:

  1. Have to say, I admire the pragmatism of the Board. Dr Thaler, where's the problem? Name yourself as inventor!

    The thought comes to mind "No harm, no victim? Then no crime."

    How this line plays out in jurisdictions outside the EPC is their problem, not that of the Legal Board of Appeal.

    ReplyDelete
  2. "Unlike the courts and patent offices, however, the willingness of the patent community to take Dr Thaler and his "imagination engines" seriously is rather embarrassing."

    I am not sure this is fair. I have not seen any widespread feeling in the patent community that DABUS definitely is an inventor. For many, the question is not whether this particular AI is capable of inventing, but whether an AI could in principle be an inventor, if not now, then in the future. Dr Thaler's various patent cases and appeals provide a convenient opportunity to discuss it. Even if DABUS is not capable, I do not think it is embarrassing to consider and discuss the issue in advance, and consider any changes in the law that might be required, instead of waiting until we have an AI invention that is somehow confirmed as genuine before trying to sort it all out.

    ReplyDelete
  3. AI has the particularity of being purely artificial and is anything but intelligent. When realising what is at the source of AI it becomes clear that it cannot be intelligent. AI boils down to predict a result by using data from the past by either crunching big amounts of data or by predicting a possible outcome. As the data itself or the prediction is based on the past, it is difficult to envisage how anything new can be created.

    A machine can be good at playing go or chess, but the same machine cannot play chess when it has been programmed to play go and vice versa. A human being can, even if it less than perfect.

    When Mr Thaler comes up and tells us that his algorithms provide credible evidence for a "virtual afterlife", and the religious concepts of salvation and damnation, this is bordering on charlatanism. Trans- and post-humanism are philosophical theories, but have nothing to do with scientific reality.

    At the moment AI is a big hype and a good means to get money from the state. Governments are eager that their country should not be left behind, should the hype become reality. But like any hype, AI will die and then subsidies with die with it.

    We might be far from patent law, but the pragmatic approach of the BA is to be welcomed. The position of the German federal Patent Court is less clear, allowing Mr Thaler’s representatives to even claim a half victory.

    When looking at the two specific cases conclusions are easily been drawn. The first one, the can with the fractal surface, is barely inventive. The second one, the flashing light is on the limit of sufficiency. In any case nothing worth getting excited about.

    The number of patents on AI might be rather limited. Without the training data no sufficiency. As data is the gold of modern times, I cannot see many firms wanting to disclose training data. The whole discussion is therefore rather academic. There are more important things to discuss than some ideas from Mr Thaler.

    I fully agree that “the willingness of the patent community to take Dr Thaler and his "imagination engines" seriously is rather embarrassing."

    ReplyDelete
  4. The technology may not be able to replicate human intelligence yet, but it's naïve to suggest it's impossible that it could get there one day. To consider human intelligence above replication seems vain, we are not magical in any way - just because we don't fully understand something now does not mean that will always be true. In some ways, we are ultimately exceptionally complicated machines, but machines nonetheless.

    I completely agree with the sentiments above that AI is, at the moment, subject to a lot of hype and there is much lauding of developments as if they are closer to strong AI than they really are. I do not, however, think the entire field of study can be simply dismissed out of hand, nor do I think that its future is certainly doomed.

    ReplyDelete
  5. Attentive, one interesting area is the use of "AI" in the search for new molecules with a useful medical effect. Patent attorneys in-house with "Big Pharma" are see in which way the wind is already blowing. At least one such attorney posts on the US blogs to say that we are getting closer to the day when credit for the contribution to the art should be attributed to the AI.

    Thaler is using his DABUS cases to rehearse how the law can handle such cases.

    The Legal Board has shown us with quite some elegance how to manage such cases pragmatically and fairly under the EPC. Its road map doesn't get you anywhere though, in the inventor-obsessed USA. There, they have with their current patent statute, a looming AI problem. But it's their problem, not ours, thank Goodness.

    Remember the concept of "enhanced novelty", anybody, to manage "secret" prior art? Will there now be some awful international "compromise" to manage cases in the future when the inventor is deemed to be an AI?

    ReplyDelete
  6. I'm also not sure it is true that the "patent community" is on the whole taking Dr. Thaler seriously (embarrassingly so, or otherwise).

    I am one attorney who is more than a bit sceptical of the claim that this particular invention is truly invented by an AI. From the conversations I have had, I do not think I am necessarily in the minority.

    I understand the point about forcing what may be (sooner or later, depending on who you ask) a genuine question of principle as to whether the patent law can accommodate AI inventors. The question may well require serious consideration. However, to get the best answer to that question in terms of the legal rigour applied to the answer and the legitimacy of the answer as authoritative (even to those who may disagree) I think it may be counterproductive to have it answered against a background where many are with good reason sceptical about the background facts which the judges cannot investigate but must assume.

    There was a dissenting judgement in the Court of Appeal DABUS case which shows it is a difficult question. Answering that question against a background of hotly-disputed but assumed facts is distracting, and probably creates some bias. I'm sure the judges did their best but I think the way this question has been brought to courts and offices around the world is a little regrettable.

    ReplyDelete
  7. I completely agree with Rose Hughes’ lucid conclusion :
    « As a final observation, it is a shame that the present debate over AI inventorship is based on questionable science (IPKat). Dr Thaler claims that his algorithms are not just capable of inventing, but that they are also capable of sentience and near-death experiences. As long ago as 1996, Dr Thaler also claimed that his algorithms provide credible evidence for a "virtual afterlife", and the religious concepts of salvation and damnation. The Board in J 8/20 stressed that the appeal did not relate to whether an AI, or DABUS in particular, is capable of inventing. Unlike the courts and patent offices, however, the willingness of the patent community to take Dr Thaler and his "imagination engines" seriously is rather embarrassing. »

    The irony is that Rose Hughes’ conclusion is in line with Dr Thaler’s own admissions, published on his website « Imagination Engines Inc » in 2019, in the FAQ section (apparently suppressed in the current version). Here is the literal quote, which tells volumes about Dr Thaler’s opinion about peer review and what « science » means.
    FAQs
    2. How come Thaler hasn't written a ‘landmark’ paper on DABUS?
    In short, Thaler has written a landmark patent on DABUS and submitted it to totally unbiased subject matter experts (a.k.a., patent examiners) for approval. Besides, IEI is a business and not formally a part of academia where professors are paid to spend most of their time writing papers.
    Then again, our founder has written peer-reviewed papers on DABUS that are purposely a bit cryptic considering the related patent that was then in prosecution.

    Also of note regarding the credibility of Dr Thaler’s claims is the EPO’s refusal of his application EP3092590 which relates to the DABUS system. This application has been rejected inter alia for lack of clarity for the reference to such terms as « thalamobot » and »imagitron » deemed devoid of technical meaning.

    Dr Thaler’s claim that DABUS is able to invent without any human input somehow reminds us of that famous story of the Automaton Chess Player, in which an expert chess player of small stature was hidden.

    As for the use of IA in pharma to speed up drug discovery, it has been in commercial use for some time. The implementation has nothing to do with Dr Thaler’s claims as it relies on a suite of highly specialized, well documented tools : chemical space exploration for identification of molecules matching a certain target, multiparameter optimization, synthesis planning, each requiring extensive expertise and data resources. Interestingly enough, there may be no need to mention the use of AI in in an application covering the output of the process.

    ReplyDelete

All comments must be moderated by a member of the IPKat team before they appear on the blog. Comments will not be allowed if the contravene the IPKat policy that readers' comments should not be obscene or defamatory; they should not consist of ad hominem attacks on members of the blog team or other comment-posters and they should make a constructive contribution to the discussion of the post on which they purport to comment.

It is also the IPKat policy that comments should not be made completely anonymously, and users should use a consistent name or pseudonym (which should not itself be defamatory or obscene, or that of another real person), either in the "identity" field, or at the beginning of the comment. Current practice is to, however, allow a limited number of comments that contravene this policy, provided that the comment has a high degree of relevance and the comment chain does not become too difficult to follow.

Learn more here: http://ipkitten.blogspot.com/p/want-to-complain.html

Powered by Blogger.