Two recent decisions in South Africa and Australia have accepted the fanciful thesis that an AI could be an inventor. I do not intend to comment on these decisions here, I simply want to try to understand how such a far-fetched (and useless) idea could gain so much ground in such a short time.
Let’s say it right away: although this thesis has always seemed fanciful to me, even to the point of making me smile, I thought it was perhaps necessary to ask (more) why it might seem appealing, instead of criticising it without having pondered it for a moment. Try to listen to her, not just hear her, in order to understand her, and thus understand her success (although it is all media, first and foremost).
This “meditation” will have led to two lines of thought.
Man is ‘enveloped’ by technology
First of all, Man, in his relationship with technology, often fails to understand the whole of the technical phenomenon. In other words, our world is enveloped by technique, so that we can understand precisely how each device works, and we lose our direct relationship to the world [1].
Thus, in order to write this article, my word no longer passes through the written hand (manu-scrit), but passes through the intermediary of a sophisticated machine which will have implemented numerous algorithms to do so.
It seems to me that the same applies to the inventor, on a more or less large scale. Just by using a computer, one does not know precisely what is going on in the “black box”. Does this mean that all inventions that use computers are not inventors? No. That depending on the sophistication of the computer a different position should be taken? I don’t think so either. The reply is that AIs are much more powerful. Yes, but do they have the capacity to decide beyond the instructions they are initially given? No. Moreover, the question of the material used by the inventor and the margin left to be inventive has already been raised by the jurisprudence. In any case, this reflection on AI taken to its extreme tends to raise questions about the very existence of our inventor. Are there still any geo-finders?
The epinal image of Géo trouvetout persists
Paradoxically, it is undoubtedly the image of Géo trouve tout, of this genius inventor, that is at the origin of the success of the DABUS affair. Why has this idea, however burlesque it may seem, been able to gain so many followers in such a short time? Probably because it makes shortcuts that all lead back to the mythology of patent law. Since ancient Greece, the craftsman and his technè (τέχνη) have always had the greatest difficulty in finding a place for themselves in front of the poet and his creation (the poet was seen as the only creator before the development of monotheistic religions) [2]. Yet it was not until the twilight of the eighteenth century, at the dawn of our modern patent law, that the inventor was glorified as the emblematic figure of Genius [3]. One historian has shown that this was the result of a desire for recognition of a new social group rejected by the traditional “Academy”, i.e. by scientists [4].
It seems to us that the emergence of this figure can be linked to the advent of machinismo following the Revolutions, or even to the source of the advent of the Modern Age (in the fifteenth century with Gutenberg and the printing press, Copernicus, etc.).
In any case, as early as Article 1 section 8 paragraph 8 of the American Constitution, the Founding Fathers of the United States of America declared “Congress shall have Power […]. to promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries.” And the patent system across the Atlantic has until recently remained committed to this emphasis on the inventor, to the detriment of the applicant. Thus, from the end of the 18th to the beginning of the 20th century, the figure of the inventor as we know it was born: a single person, endowed with an extraordinary intelligence, who, alone in his cellar, advances humanity through his mind. Bergson thus saw invention as the characteristic of human intelligence, the most elusive aspect of it, making the inventor a genius [5].
This figure of the inventor still continues today the common meaning of the term “inventor”, and can, at times, continue to animate the legal meaning of this term, as understood by patent law [6]. However, the mythology, which the Greeks did not believe in themselves [7], continues to captivate the uninitiated: the inventor, this extraordinary character, would be fundamental in the patent law system and should necessarily be the owner of his invention. Hence two (erroneous) ideas put forward to justify the recognition of the AI as an inventor: recognizing an inventor other than the AI (when it could supposedly invent alone) would be tantamount to not recognizing the “right” owner, it would diminish the inventor’s moral right (which is in reality only a right of paternity intended to follow the chain of rights) and, above all, it would prevent the patentability of the inventions in which the AI would have participated.
However, any patent law practitioner knows the reality: the applicant is the owner of the property right while the inventor is, ab initio, excluded, being presumed that his right has been assigned to the applicant from the start [8]. This is the famous distinction between the first-to-invent and first-to-file systems. In such a system, adopted by all countries in the world, the inventor has only two (marginal) rights: the right to claim ownership of the invention if he proves that it has been stolen from him, and the right to be named inventor. Therefore, in the end, it does not follow that the arguments of the “defenders” of the AI inventor hold water, if one remembers that: an AI never invents alone and that the claimant (its owner) will necessarily be the owner of the invention in which its AI participated. In other words, recognizing an AI as an inventor requires twisting the legal system (notably on the issue of legal personality), creating unnecessary confusion and insecurity, for no result, since the AI will be protected anyway [9].
To conclude, this story of AI as inventor reminds us of the power of Myths, and, at the time of the bicentenary of Napoleon’s death, that “imagination rules the world” (for good or ill, one might add, as the Emperor himself demonstrated).
Matthieu Dhenne
Avocat (Paris Bar)
[1] V. M. Heidegger, Parménide, Éditions Gallimard, spéc. p. 140 et 141.
[3] Platon, Le Banquet, Les Belles Lettres, Universités de France, 1992, no 205 b), p. 58 et 59.
[4] Even if Bacon had already sketched this glorification of the inventor in the middle of the 17th century. Bacon, Novum Organum Scientiarum, Apud Adrianum Wijngaerde et Franciscum Moiardum, 1645 (1620), no XXXVI, p. 40.
[5] L. Hilaire-Pérez, L’invention technique au siècle des lumières, Albin Michel, 2000, pp. 147 et s.
[6] H. Bergson, L’évolution créatrice, PUF, 1957, p. 165.
[7] The legal “burden” of a term often leads to different meanings between common language and legal language, as in the case of treasure, which, according to Article 716 of the Civil Code, is “a hidden or buried thing over which no one can justify his ownership, and which is discovered by pure chance”. “See P. Amselek, “Philosophie du droit et théorie des actes du langage”, in Théorie des actes du langage, éthique et droit, ed. P. Amselek, PUF, 1986, p. 109. Not to mention the excess of the signifier (sign) and what it designates (the signified). The signifier “invention” actually covers a large number of very different realities. See G. Deleuze, Logique du sens, Les Éditions de Minuit, Critiques, 1969, p. 63.
[8] See M. Heidegger & E. Fink, Héraclite. Séminaire du semestre d’hiver (1966-1967), Gallimard, 2017, où Heidegger rappelle que les grecs eux-mêmes ne croyaient pas à leurs mythes.
[9] See J. M. Mousseron : Le droit du breveté d’invention. Contribution à une analyse objective, LGDJ, Bibliothèque de droit privé, t. 23, 1961.
[10] See M. Dhenne, Intelligence artificielle et droit des brevets, Propriétés Intellectuelles, n° 78, p. 23.