With a valuable contribution from Lorenzo Attolico, partner of the NCTM Law Firm, we are inaugurating a research and in-depth analysis on artificial intelligence:

the incredible potential of human and technological development that today is facing us in the light of the fourth industrial revolution, the opportunities for collaboration between the Canadian research world and the European one that derive from it, but also the questions that arise in terms of rights, of protection of intellectual property in the new economic paradigm that is emerging.

In the article by Dr. Attolico, we are familiar with the possible solutions for the protection of intellectual works produced by artificial intelligence, asking ourselves a question: can robots be considered authors of an intellectual work and therefore holders of relative economic rights?

 

The AI ecosystem in Canada

 

The AI scenario is dynamic in Europe, and even more so in North America, where the digital revolution started. We are talking about Silicon Valley USA, a crucial global research and development spot in the IT field. It has not gone unnoticed, however, as the industry giants - Google, Facebook, Microsoft, Samsung, Uber - and the biggest names in technology are investing millions of dollars in research laboratories in Canada, where academic research and innovation are strongly connected and the attraction to researchers from all over the world is therefore very strong. The Canadian government continues to support innovation and has announced an additional 4 billion dollars in scientific funding for the coming years. Thanks to this public support, Canada has come to be the third artificial intelligence ecosystem by IA's number of experts, a world leader in robotics and machine learning development, at the center of research that tracks new ways to extract big data with implications for all sectors: health, automotive, domotics, aerospace.

In the last 5 years the increase in agreements in the AI field has reached 50%, with a constant presence by Canadian investors and a stable share of about 40% of international investments.

 

 

From 2017 to 2018, there was a 28% increase in the number of active startups related to the AI, with almost 650 startups active in all major Canadian cities.

 

Toronto has the largest concentration of AI companies in Canada, a solid academic history, an established technology industry and takes advantage of its proximity to key sectors, such as finance, and the constant government support for industrial development of Artificial Intelligence.

 

The Montreal artificial intelligence cluster is a world reference center for artificial intelligence research, with the largest concentration of AI researchers in the world.

 

Even Vancouver is emerging as a unique center of the west coast for artificial intelligence. Despite not having the concentration of resources or research skills that characterize other Canadian clusters, it is developing a strong ecosystem of startups and important international connections.

 

The hope is that the European Union and Canada, partners that share the same values, the same principles, the same vision of the world, will strengthen closer cooperation in this area, in the field of theoretical and applied research, for the creation of new collaborations and joint research projects, technology development and technology transfer.

 

 

Legal Profiles of Intellectual Works Created By Artificial Intelligence*

 

LORENZO ATTOLICO**

 

Summary: 1. Traditional and new authors of intellectual works. -  2. European Parliament’s Resolution of 16 February 2017 with recommendations to the Commission on Civil Law Rules on Robotics (European Parliament 2017). - 3. The framework in force in other States - 4. The possible solutions to the protection of intellectual works created by an IA and the ownership of related rights. - 4.1. Economic rights - 4.2. Moral rights. - 5. Third party liability for the use of intellectual works created by a IA.  – 6. Conclusions.

 

1. Traditional and new authors of intellectual works

 

From the beginning of the history of humans the creation of intellectual works has always been attributed to human beings, who were, indeed, the only ones who could consciously create ex nihilo works having artistic features.

 

It is pointless to make a reference to the first prehistoric graphics, which, even if in a roughly, poorly evolved and usually monotonous way (hunting scenes only were represented), prove to be the first proofs of the artistic verve and need intrinsic to humans.

Literary, architectonic, theatrical and musical creations followed over time figurative and sculptural works of art, until, in the past century, works resulting from new technologies, such as photographic, cinematographic works and, finally, computer programs, were reached. 

All these works, as said, had something in common and, namely, the fact that the originating entity was a human being, therefore endowed with intellect.

For this reason, when copyright policy concerns were raised, the author of the work was naturally identified as recipient of rights on the intellectual work.

In this sense, also our law on copyright (Law 22 April 1941, n. 633), which provides in art. 6 that “Copyright shall be acquired on the creation of a work that constitutes the particular expression of an intellectual effort”, made express reference to the human author.

Thus, after all, could not be anything other, given that intellectual works could not be created by entities other than human beings.

However, this approach is changing fast.[1]

The first case, in order of time, which has characterised the beginning of a new hera, is represented by the issue relating to the ownership of rights on a photographic work realized, such as selfies, by a monkey, and, indeed, an animal. The case is well known.

The professional photographer David Slater, when in Indonesia for making a photographic service of a group of monkeys left a camera unattended. And that’s when a macaco grabbed it and started to take, unintentionally, a series of photographs, among them tens of selfies. Two of those came out, by chance, perfectly.

It is then that he rightly asked himself, who were the owners of the rights of economic use in the picture: the photographer, the owner of the camera and, by its saying, who set it up in order to make happen what effectively happened, or, nobody, given that the macaco could not be considered copyright holder.

The happening reached the American courtrooms, where, at least at first instance, the photographer has not been recognized as legitimate holder of any right.

The decision was delivered on the ground that – important also for works created by machines - worth of protection is only a creation resulting from an intellectual work („fruits of intellectuale labor “ that „ are founded in the creative powers of mind“).

It was not, therefore, in essence, thought possible to attach to a person different from the human being the quality of author of an intellectual work.

Recently, it was then found out that the querelle has been settled by negotiation through recognition, by the photographer, to an animal rights association of a percentage of revenues from the use of the picture.

However, in the present case, the picture was taken unintentionally by the macaco.

Different is, instead, a situation in which an entity, also when other than human, creates consciously an intellectual work.

This hypothesis is possible since when artificial intelligence has been introduced in the technological world.

 

Artificial intelligence (IA) can be, in essence, defined as a simulation of human intellect by a computer or other machines, such as robots.

The technological evolution is thought to be bringing up the creation of autonomous and intelligent machines surely ready to take decisions independently of human creators or users in order to be able one day to perform our intellectual capacities, and even to surpass them.

It is well known that a robot created autonomously a series of musical works which will soon be part of an album to be commercialized, as well as that two robots have just completed respectively the drafting of a script and an out-and-out literary work.

These have all been created independently thanks to the inputs which have been given to a IA machine from the respective creators, inputs which, nevertheless, as it happens for humans, have represented the starting point of the creation.

The impression is that this kind of new works would become always increasingly common.

In fact, it will be perfectly normal to take advantage of and enjoy intellectual works created by IA machines.

In other words, human beings will only limit themselves to enter disaggregated data in the machine, which, as previously indicated, will only form a platform, considered as the cultural background of a human being, over which the IA will create a new work.

It is, therefore, correct and appropriate to be asked whether robots, representing a new category of legal entities, could be held authors of an intellectual work and holders of relative economic rights.

 

2. European Parliament’s Resolution of 16 February 2017 with recommendations to the Commission on Civil Law Rules on Robotics (European Parliament 2017)

 

On February 16^th, 2017, the European Parliament submitted to the Commission some Reccomendations (A8-0005/2017) dealing with civil laws on robotics.

Indeed, the Parliament intended to inform the Commission on the necessity to verify the suitability of actual civil laws to govern situations which will necessarily increasingly ‘go live’ due to the activity of machines equipped with artificial intelligence and more particularly robots.

Indeed, we are certainly at the entrance to a new industrial revolution and, therefore, legal systems, particularly the European, should update in time in order to be found unprepared.

As is well known, the Commission is free to follow the suggestions of the Parliament, but, in case of non-compliance – which happens relatively rarely – has to give a reason for it.

It is well possible, therefore, already in this initial phase of the European legislative procedure, to comment on the content of the parliamentary document, having a reasonable expectation of a final decision measure along the lines suggested. 

The document of the European Parliament deals actually with different legal aspects and in particular with the impact of machines equipped with IA on the working activity, as well as the fundamental topic of attribution of responsibility deriving from acts committed by those machines and, finally, of a topic apparently less important, but in fact very actual, and, namely, the applicability of copyright laws, to the IA world.

In other words, the Commission raises the issue of whether a robot can be considered author of an intellectual work, becoming holder of related economic rights and, moreover, whether it can be held responsible in case those works result, later, plagiarism of works created by third parties.

In relation to the first of the topics of above, the Parliament invites the Commission to develop criteria to define a “intellectual creation on its own” by a machine, suggesting – it seems – substantially the creation of a new genus of legal entity, owner of the intellectual property: an electronic entity.

In terms of responsibility, conversely, it does not seem that the Recommendations tackle specifically the issue concerning immaterial goods, although, on the robotic issue in general, the topic is considered to be extremely sensible and important.

 

3. The framework in force in other States

 

The concern of the European institution to govern the topic seems appropriate also in the light of the fact that some States (Japan, South Korea, China) have already started to prepare a specific framework.

In the United States, although the Copyright Law Act of 2011 does not involve expressly the recognition of copyright ownership on intellectual works created by a machine, the question was asked whether those works could be covered by Copyright in some way.

In this respect, recently the US Copyright Office stated that it will provide for registration of an intellectual work only if this has been created by a human being (“will register an original work of authorship, provided that the work was created by a human being”).

However, the US Copyright Office was also careful to provide a list of non-protected works among which express reference is made to works created by a machine and resulting from an automatic or casual realization devoid of any human creative intervention (“works produced by a machine or mere mechanical process that operates randomly or automatically without any creative input or intervention from human author”).

As regards Japan, a Commission established by the government in 2016 stated that the law on copyright does not protect works created by IA, but the government itself seems to intend to research some form of protection for these in the current year. Japan itself, conversely, attributes to an invention created by a IA the possibility to be patented.

Finally, in the UK the Copyright Designs and Patent Act of 1988 provides that the ownership of copyright on a work created by a machine belongs to the entity (human being) which has organized in some way the functions of the machine itself, so that this could create the work in question (“In the case of literaly, dramatic, music or artistic work which is computer-generated, the author shall be taken to be the person by whom the arrangements necessary for the creation of the work are undertaken”).  In the same Act, inter alia, the definition of “computer generated work” is given, which is, essentially, the work created by a computer without any human intervention (“the work generated by computer in circumstances such that there is no human author of the work”).

As underlined, other legislations tackled in different ways the issue of the eligibility for protection of intellectual works created by machines, and came up with various solutions.

The English legislation seems to have mapped out the most interesting path, which, in fact, recognizing a value to those works – conclusion which, in the opinion of who writes, can now be considered essential –, attributes the ownership of the rights to the human being which gave to the machine the necessary tools for the creation.

Surely this can be one of the paths to be followed also in our and in the European legislation, but surely not the only one and maybe also not the most convincing, because – however this will be mentioned below – in the evaluation of the present case it won’t be easy to understand whether there has been or not the implementation of sufficient tools to give rise to a certain creation.

 

4. The possible solutions to the protection of intellectual works created by an IA and the ownership of related rights.

 

Until today in our legal system, according to art. 2576 c.c. and, how already recalled, art. 6 Lda, the ownership title of copyright consists of the creation of the work as a particular expression of the intellectual work.

 

This means in essence that:

• owner of all rights on an intellectual work is the author of the work;
• the author, considered the precise reference to the concept of “intellectual work” has to be necessarily a human being;

Now, leaving aside the delicate evaluation of the possible creation of the “electronic entity”, as a new legal entity, I have the impression that the such setting can’t be sic et simpliciter used for intellectual works created by a IA.

In fact, there is no doubt that the IA, as a machine, whilst be considered author being the creator of the work, can’t be held addressee of economic rights, and even more moral rights.

It will therefore be appropriate to provide for a legislative amendment, ideally at the European level, aimed at governing this legal situation, which allows to reserve to a “classical” legal entity (no matter if natural or legal person) economic rights, and if possible, moral rights on the intellectual work created by an IA.

I will try, in the following, to imagine in which way this new situation should be governed.

 

1. 1. Economic rights

 

The creator of an intellectual work, as is known, is the exclusive owner of the economic rights and has the right to benefit from the revenues from its exploitation.

As the great Mario Fabiani foresaw back in 1933 “in the present landscape we must rethink the principles and the rules governing copyright law”.

To this end, I think three scenarios can be drawn up:

1. Owner of the economic right under copyright on a work created by a IA is the entity which invented the machine;
2. Owner of the economic right under copyright on a work created by a IA is the entity which in some way set up the functions of the machine itself, so that this could create the work in question;
3. Owner of the economic right under copyright on a work created by a IA is who, being the owner of the machine and regardless of who set it up, initiated the economic exploitation of the work itself.

The first scenario I suggested for the sake of completeness does not seem to deserve consideration.

The entity which merely created the IA (think of the creator of a software) cannot in any way demand to claim to hold the rights in question, not having organized its specific functions which allow the creation of a certain work.

Think of the case of an author of a software who allows the realization of works of design. There is no doubt that the author of the work of design is the person who de facto creates it.

Now, the second scenario is slightly different, but the core issue remains that it is not sufficient to create a software to be granted the ownership of rights of economic use of the works created there through. 

The second scenario is particularly interesting and, as already stated, was taken into consideration by the Italian legislation.

However, it remains hard to understand on a case by case basis whether, and eventually which, “arrangements necessary for the creation”, as defined in the UK Copyright Act, have been provided.

In fact, it is hard to determine with a reasonable degree of certainty whether a IA created a work (and not another) only on virtue of the making of certain arrangements.

Think of a robot able to create a musical work. How can you be sure that, once the work is created, this was the direct result of human intervention in the upload of certain musical tracks or in the giving of inputs.

Therefore, in the absence of such evidence, I don’t think the “uploader” entity can be qualified as author.

The assessment of the case would change radically when the technique could permit such an evaluation and allow a certain attribution of the merits of the work to the “uploader” of the arrangements for the creation, who could be held to be the author by the interpreter. 

Finally, the third scenario seems to be the most reasonable to me.

Therefore, previously it is to be clearly ascertained that the so-called electronic entity cannot be considered the author of an intellectual work, nor the owner of the economic rights on the work, and that it remains impossible to consider the uploader of content the creator of the same work. It is subsequently evident that the rights referred to above belong to every person who exercises the ownership of the rights on the assumption that these have been acquired lawfully, i.e. not as the result of an assignment (which can’t be conceivable since the real author is the electronic entity), but rather because the user is also the owner of the creator, i.e. the IA machine.

In other words, the owner of a IA machine, who created an intellectual work, should be considered also the owner of all economic rights on the work itself.

In this last case, it will be irrelevant to assess who created the content. Therein a substantial difference lies between the third and the second solution.

In this respect, principles applicable to works on commission will come to the aid for which, as is well known, rights of economic use arise under the original title by the costumer[2].

The scenario where the owner of a IA machine activates it to achieve an intellectual work, can be easily assimilated to a contract for services.

 

1. 2. Moral rights

 

The need to adapt the actual rules to the technological news seems to be really ineradicable, in particular with reference to the issue of regulating moral rights under copyright on an intellectual work created by a IA.

If, as already said, it is possible to identify the holder of economic rights under copyright, on the other hand it seems very hard to determine the holder of moral rights under copyright.

In fact, those rights belong exclusively to the author and are unavailable.

It follows that no other entity, if not, indeed, the author of the work, can be considered holder of those rights.

In your case, since the author of the intellectual work is an entity which cannot be holder of any right, still less personal rights, which are obviously owed to human entities, is quite evident that no entity can be entitled to moral rights under copyright.

Therefore, from the intellectual works created by a IA no moral rights will descend and this counts as an exception to the rule set up in artt. 20 e ss. Lda.

 

5. Third party liability for the use of intellectual works created by a IA. 

 

At this point, it only remains for me to address the topic of the liability for damages to third parties caused by an intellectual work created by an IA.

For instance, take into account the responsibility in case that work result to have been plagiarized from another pre-existing work. 

However, also in that case the liable entity shall be identified.

On this point, great help is provided by the Recommendations mentioned in paragraph 2) which tackle the issue and suggest some solutions which appear to be rather reasonable.

Firstly, it should be made clear that the liability shall be attributed to a human entity and not to a robot and that it won’t be possible to restrict in any way nor the kind and the entity of damages which can be compensate, neither the forms of compensation which can be offered to the injured party. Therefore, from the mere fact that the damage is caused by a non-human subject, it is affirmed that the future legislation primarily will have to decide whether “adopt a system of objective responsibility or risk management”.

In the first solution, in fact, the mere evidence of the incurred damage will be required and the identification of a causal link between the functioning of the robot and the damage suffered by the injured party. Whereas, the adoption of the second solution will not require the identification of the person acting with negligence in its responsibility on an individual level, but upon the person who “in certain circumstances is able to minimise the risks and address the negative impact”.

In the Recommendation mentioned above these possible disciplines are actually suggested not expressly for the topic linked with intellectual works, but generally as an aspect connect to AI.

It follows that the mentioned principles will necessarily have to be adapted to the intellectual property world generally and the copyright world in particular.

In this context, it seems to me that among the two suggested solutions the Risk Management System should be preferred. 

The owner of the economic rights on the work, indeed, has the possibility to minimize the risks and, on the other hand, tackle the negative impact.

In order to minimize the risks, the inspection of an expert will have to be verified – I mean before the commercialization – in order to determine whether a work has plagiarized another and, once in commerce, in the event of a dispute it will possible to withdraw it from the market. 

However, the principle results to be in line with what has already established by case-law. Namely, the publisher will be held responsible with the author of a work which ultimately resulted to be plagiarist.[3]

The subject who benefits for the revenues of the economic exploitation of the work, is deemed to be able to face also the predictable negative impact of a plagiarism dispute.

 

6. Conclusions

 

To conclude, it seems possible to say that the future is now. Intellectual works, but not only, are already created autonomously by the IA, which will be always more.

For once, therefore, the regulation can’t wait.

In fact, topics such as ownership of rights and responsibility to third party claimant should necessarily be solved, in short times indeed, to keep abreast with leading-edge technology.

One day robots will be indispensable, but that day we will have to be ready with a rigorous discipline governing their use and its consequences.

Indeed, it won’t be possible to rely on the Three Laws on Robotics coined by Isaac Asimov[4]. 

 

 

Bibliographical references

 

AUTERI P. FLORIDIA G. – MANGINI V. – OLIVIERI G. RICOLFI M. SPADA P., Diritto industriale. Proprietà intellettuale e concorrenza, Torino 2012.

FALCE V., La modernizzazione del diritto d’autore, Torino, 2012.

MAZZEI G. – PICCINELLI G. – TISCI A., Digital properties and digital consumers – Nuovi diritti e nuove tutele, Napoli, 2012.

CROPANESE R.,  LO FOCO M. Il diritto d’autore – storia evoluzione regole futuro, Torino, 2013.

SPEDICATO G., Interesse pubblico e bilanciamento nel diritto d’autore, Milano, 2013.

SIROTTI GAUDENZI A., Il nuovo diritto d’autore. La tutela della proprietà intellettuale nella società dell’informazione, Rimini, 2014.

M.L. MONTAGNANI – M. BORGHI, Proprietà digitale, Diritti d’autore, Nuove tecnologie e digital rights management, Milano 2006.

G. ZICCARDI, Nuove tecnologie e diritti di libertà nelle teorie nordamericane.

G. GHIDINI, Profili evolutivi del diritto industriale. Proprietà intellettuale e concorrenza, 2° ed., Milano 2008.

A. SIROTTI GAUDENZI

 

 

*Originally Published On Linkedin Pulse

**Head of IP Department at NCTM Studio Legale