Report: EPO conference - Patenting Artificial Intelligence

On 30 May 2018, the EPO held a ‘first of its kind’ (as it was called by one of the EPO officials) conference on ‘Patenting Artificial Intelligence’. The event reportedly sold out one week after registration opened, showing the tremendous interest from professionals and industry in the field.   Katfriend Olga Gurgula (Lecturer in Law at Aston Law School and Visiting Fellow at Oxford Martin School, University of Oxford) was on hand to report back to the IPKat readers, and some moral musings from Gwilym Roberts on these issues will follow shortly in a separate post.

"This timely event coincided with the confirmation by EU Ministers that AI will become one of the most significant mechanisms for boosting the competitiveness of the EU and the wellbeing of its citizens.

The main thread connecting all the sessions was the common understanding that AI provides an immense opportunity for innovation. Moreover, the patent system must respond adequately in order to ensure that it does not stifle but enhance such innovation. The main goal of the event was therefore to facilitate a discussion about the challenges in patenting AI.

The first two keynote speakers analysed AI as a driving force of the 4th Industrial Revolution (4IR). Clare Dillon, a technology evangelist, gave an insightful introduction into the topic and shared her vision on the formula of success for AI, namely the combination of ‘Technical innovation + digital culture + economic viability’. Her presentation was followed by another keynote speaker, Yann Ménière (EPO Chief Economist), who highlighted the dramatic increase in patent applications related to smart objects within the last several years. A third keynote speaker Koen Lievens (EPO Director) explained how the EPO examines patent applications in the area of AI.

The remainder of the first half of the day was dedicated to general strategies for protecting AI inventions and the challenges in drafting patent applications for AI-related inventions. It touched upon how the EPO conducts its examination of the subject matter, including its two-hurdle approach and use of the developed case law in this regard. It was also noted that the EPO has various mechanisms in place to support improvement to its activities in this area, including its annual revision of the guidelines for examination of computer implemented inventions (CIIs), the revised version of which will be available online in Autumn 2018. Panel experts also shared their experience and provided various examples on how they overcame challenges encountered during the examination process of AI-related patent applications.

The second half of the day was all about post-grant aspects of AI patents and ownership in different jurisdictions, including Europe, the US and Japan. The concluding session addressed controversial ethical aspects and considerations associated with patenting AI related inventions, including such issues as AI and IP related to consumer protection and security, as well as who is entitled to a patent for an artificially generated invention and who should benefit from such inventions. An interesting discussion was generated by a presentation of Dr Noam Shemtov (Queen Mary University of London), who shared his thoughts in relation to the autonomous inventions, and whether the current standard of the assessment of obviousness bench-marked against the ‘person skilled in the art’, i.e. human, is adequate, or whether it may require some adjustments to scenarios involving autonomous invention processes.

In a lively Q&A, one of the participants, Gwilym Roberts from Kilburn & Strode, raised a fundamental question in relation to AI inventions. In particular, as technology develops and AI begins to invent more and more, is there a risk that the owners of AI will begin to dominate the IP sphere? In other words, big companies or countries that invest in AI now could soon have a super-monopoly in this area. Therefore, the question is whether the current policy is well equipped to deal with this problem at present. Dr. Shemtov suggested that perhaps competition law, rather than patent law, is better suited to deal with this problem.

The speakers and audience shared a range of other interesting thoughts and ideas in relation to AI, which were seen as crucial to consider for enhancing the patent system in this area. These included, for example, growth of ‘black box’ AI technology, which is becoming self-explaining. It is therefore now possible to understand how it works and takes decisions, which is important inter alia from the perspective of Article 83 and 84 of the EPC. It was also noted that the EPO had habitually responded to software innovation by developing the CII regulations without changing the law, i.e. by patching the system. The room considered whether we can continue to patch the system in relation to patenting AI or whether more fundamental changes are required.

Olga's cat: natural intelligence

Some innovative proposals were mentioned, such as introducing a utility model to provide fast provisional protection; changing the philosophy of IP legislation from protecting inventors to protecting business; and publishing patent applications earlier than 18 months as the current period does not suit new technologies such as AI. Finally, the crucial question was raised: do we somehow allow the patentability of algorithms?

As Grant Philpott (EPO Chief Operating Officer ICT) noted in his concluding remarks, adapting the patent system designed for hardware innovation to the virtual world of software innovation has been a long and difficult process, but ultimately, Europe and the EPO have succeeded in developing a highly methodical approach to the patenting of CIIs, which is firmly rooted in EPC case law and EPO practice. However, as AI software becomes super-software, incredibly powerful and versatile, the challenges for our patent system are likewise significant. Events such as this help us to meet that change and ensure that AI and its impact on patents and patentability continues to be dealt with systematically and in line with the laws and needs of stakeholders. Only by doing so can our system provide the stability and predictability that the industry needs for its business models, and further support economic growth of Europe.

To conclude, the 4IR is happening right now, and AI plays a key role in these dramatic changes, impacting the scale and the speed of innovation in every industry. Examples of AI in emerging technologies, which would have doubtless been seen as science fiction less than a decade ago, are now becoming part of our everyday life. As one of the speakers noted: ‘It is not a myth, it is already here’. Therefore, a clear vision on how to boost the development of AI, while maintaining ethnical aspects in place, is a fundamental issue that requires urgent decisions involving all stakeholders and members of society, and the EPO’s initiative in calling this conference is a welcome step in the right direction."
Report: EPO conference - Patenting Artificial Intelligence Report: EPO conference - Patenting Artificial Intelligence Reviewed by Eibhlin Vardy on Wednesday, June 06, 2018 Rating: 5


  1. The slidepacks for this conference are now available from the EPO, with video apparently to follow.

  2. I would like to add that Russ Abbott wrote an interesting article a couple of years ago on this topic in the Boston College Law Review

    Volume 57 | Issue 4 Article 2

    I Think, Therefore I Invent: Creative Computers
    and the Future of Patent Law

    Ryan Abbott
    University of Surrey School of Law,

    He comes out with a favourable opinion on "Machines as Inventors": it is clear that the US patent system, however, would require some modification while the European one looks easier...

    Stephen Potter


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:

Powered by Blogger.