Title: Artificial Intelligence and the Audiovisual World (full details here)
Date: March 30th, 2022
- Prof. Javier Fernández-Lasquetty, Professor of Intellectual Property Law at IE University. Partner, Elzaburu.
- Prof. Giuseppe Mazziotti, Professor of Intellectual Property Law at Trinity College Dublin
Moderator: Prof. Giuseppina D’Agostino, Professor of Law, Osgoode Hall Law School, York University, Toronto
Global Digital Encounters 19: Artificial Intelligence and the Audiovisual World
During, and further to the COVID Crisis, the role of Artificial Intelligence became even more central for the Audiovisual World. Its massive use in this key economic sector brings legal consequences, for IP in general, and in particular in the field of Copyright and Neighboring Rights. New business models bring ethical, practical, and economic considerations relating to AI input that are of major importance both for the audiovisual world and for the public. with unprecedented legal challenges. Which meanings might be expected from this evolution/revolution? Speakers from Europe and North America will provide a worldwide picture of the future of Artificial Intelligence in the Audiovisual World, including on re-imagining IP in this area in favor of both business and consumers/society.
Prof. Laurent Manderieux explained in his upbeat welcoming speech that the emergence of Artificial Intelligence (“AI”) permeates all activities in the digital world, particularly the audiovisual realm, which has drastically altered business models and may drastically affect Intellectual Property (“IP”) rules. As a result, FIDE and the Transatlantic IP Academy were to concentrate their discussion at this conference on the massive legal implications in the sphere of IP that arise from the development and use of Artificial Intelligence technology in the audiovisual domain. Furthermore, this subject invites to develop a reflection also on a broader legal and corporate area, on developing and increasing fundamental rights issues such as personality rights, privacy rights, and so on, connected to AI.
Moderator Prof. Giuseppina D’AGOSTINO noted that we are witnessing the pace of change and disruption in the audiovisual industries at such galloping speeds, with stakes being much higher. In her first book Copyright, Contracts, Creators: New Media, New Rules (hyperlink), she analyzed the introduction of emerging technologies and copyright implications for creators across the United Kingdom, continental Europe, and North America and from an international perspective. With respect to the audiovisual industry, she provided the example of silent picture films becoming “talkies” in the 1920s when the new sound component of talking with new technology was suddenly introduced on the motion picture scene and revolutionized, or disrupted, the movie-making industry. Consequently, this new technology brought forth renewed business models of exploitation as well as a slew of litigation in court. As of late, AI has emerged as the new bold disruptor, opening up new business models and renewed means of exploitation, but with far more complexities and possibilities, as well as unanswered questions and future litigation. AI has profoundly transformed not only the audiovisual sector, but also every area of our life in society. For instance, deep fakes have now appeared on our social network, casting doubt on even basic professional interactions. These major societal issues must be addressed not only by lawyers, but also by interdisciplinary professionals with a broad knowledge base, such as computer scientists, ethicists, as well as governments and industry players working collaboratively. In this regard, Moderator Prof Giuseppina D’AGOSTINO referred to the Report Fostering the Future of Artificial Intelligence, from the York University (hyperlink), which further discusses these issues.
1. What are the latest developments of AI in the audiovisual industry? and what are some of the main IP issues presented?
Speaker Prof. Giuseppe MAZZIOTTI pointed out that the creation of new audiovisual products, such as movies or TV series, is increasingly based on data and information that users provide to owners of content platforms. Due to the influence of online platforms, consumers no longer access and enjoy audiovisual works secretely, in their private sphere. Currently, we have a strong vertical integration of audiovisual producers who also own the supplying and distributing machines, through which films, TV series and other audiovisual products are exploited. As we have observed, not only as Intellectual Property academics but also as practitioners and consultants, the significant input coming from streaming platforms’ users is capable of challenging artistic freedom. On one hand, renowned and well-established film directors might keep away from platform productions. On the other hand, early stage film creators, with little or no bargaining power vis-à-vis large media conglomerates, might be deprived of a significant portion of their freedom to create, their role being reduced to that of simple executants. He admitted that these recent technical developments are a cause for concern, but he also finds the increasing complexities arising when it comes to contracts, due to the large number of parties involved, intriguing and interesting from a law and business perspective. The potentially conflicting relationships between authors and film producers, that characterized the film and TV industries, is now being extended to the creator/platform level, adding an additional layer of negotiation and contractual issues for practitioners.
Speaker Prof. Javier FERNÁNDEZ-LASQUETTY put the point of view that the industry works to refine how the AI algorithm determines the choice of users, in an attempt to encourage them to use the catalog as much as feasible. From the perspective of business, the implementation of AI technology has created an opportunity for independent producers that they did not have in the past, while more novelties in the AI field, such as the use of the metaverse for these audiovisual products, are still being researched. He further argued that we will witness more novelties that are not entirely tied to AI, but rather to the business model that integrates new content creators on social platforms.
2. Should we protect AI-generated audiovisual works or not? If yes, then how can we protect such output? What are the primary considerations when deciding whether to protect the AI generated outputs, and, if complexities arise, what is the best approach to resolve them?
Speaker Prof. Javier FERNÁNDEZ-LASQUETTY responded that, on the one hand, the development of audiovisual works is constrained in some ways due to the fact that human beings are involved. On the other hand, apart from the traditional players such as MGM, Universal, Paramount, some new as Netflix, YouTube and Apple TV, are entering into the production and distribution of audiovisual works. All are conscious that they want to produce movies faster and AI would be one of the solutions for that problem. Because of AI based developments (such as augmented reality or virtual reality), content can be created without the involvement of a human. It is even possible to create characters that are not human beings (using AI tools such as “this person does not exist”) As a result, we will be thrust into a new reality in which we will debate the ownership and authorship of AI-generated products in a specific manner.
Speaker Prof. Giuseppe MAZZIOTTI did not hide his skepticism on real “newness” of the dilemmas posed by Artificial Intelligence, at least in the ways many scholars have presented them so far. . As Prof. Javier FERNÁNDEZ-LASQUETTY reminded us, AI reached an unprecedented level as a concept. But at the very end of the article Authors and Machines (hyperlink), Professor Jane Ginsburg taught us that if we wish to reach the conclusion that a piece of work, even an audiovisual work, is authorless, we must provide a very rigorous legal analysis. However, the question of whether a piece of work is entirely generated by a non-human author or actor does exist. There is an interesting distinction between authorship and ownership, which matters very much in those jurisdictions. Authorship was awarded not only as a form of economic protection but also a moral protection. There has always been a distinction between the creative input and the output generated and exploded in an audiovisual format. Indeed, normally, the creative input in audiovisual works comes from a plurality of professionals and the copyright belongs to the producer. Therefore, the risk is not very much about defining authorship, but protecting those value chains that have kept the audiovisual sector relevant from a cultural and social point of view, and from a labor standpoint if we look at the performers.
Speaker Prof. Javier FERNÁNDEZ-LASQUETTY agreed on the fact that, according to the Berne convention for the Protection of Literary and Artistic Works, the current copyright system was not created for AI or machines, but for humans. The question is where to place that type of creation, as without attracting these new creations to the Berne convention, this will become a jungle. We must find a solution that lies in the best interests of everyone involved in that business. From his perspective, the solution provided by the UK law, especially in article 9.3, is not optimal, and all the other possibilities have been explored to determine how the intervention of a human being is sufficient. We must find a framework, but we must be prepared to change. Perhaps in the future, we will see two distinct types of audiovisual works: those produced in the traditional manner and those produced with a hyper-reality that perfectly replicates real life. However, it does not mean that there will be no author or creator.
Speaker Prof. Giuseppe MAZZIOTTI emphasized what the role of the law in this fast-developing framework should be. When defining authorship and ownership in different jurisdictions, we should always consider the applicable law from a political standpoint, as well as a business perspective, thinking of the types of work that a legal system wants to encourage, as that is ultimately the function of Intellectual Property. Otherwise, as Prof. Javier FERNÁNDEZ-LASQUETTY said, this kind of production exits the realm of IP and enters into other bodies of law that have nothing to do with the rationale of IP. From the perspective of the European Union, when it comes to audiovisual media policies, we should expect even more from regulators to ensure a strong incentive for human creators and human performers. If not, large media conglomerates might become even more dominant than they are now. Therefore, it is an inevitable consequence of the fact that a jurisdiction, like the EU, might place even more burdens on the conglomerate. Not to mention the reform of the Audiovisual Media Services Directive in 2018, which imposed some quotas, even on the platforms that people adore. The risk could be probably protectionism, which adds to the richness of our discussion because we see very different perspectives around the world.
3. What can be done to ensure greater diversity of content, as well as independence of smaller audiovisual producers?
Speaker Prof. Javier FERNÁNDEZ-LASQUETTY stated that, in order to operate their business, platforms provide a wide variety of content to attract enough audiences. New forms of recreation will continue in order to catch the audience’s attention. The commitment of producers and platforms will encourage them to produce everything that might be interesting to the public. The phenomenal Korean production, Squid Game, is an example. Not everything in this new marketing is wrong, it just represents the new century.
Speaker Prof. Giuseppe MAZZIOTTI added that European listeners face the challenge of unifying EU digital markets, which is a difficult thing for a variety of reasons. Market integration is being aided more by players such as Netflix and Amazon than by the law itself. Indeed, these players see the reach of the broadest and most diverse types of audiences as a business priority and have an incentive to diversify their content productions. Being able to gather data from users, platform companies are motivated to know and commercially exploit what their customers want and they do so at a pan-European and cross-border level. The only risk that this trend has triggered is an inevitable reduction in the number of movies being made digitally available. For instance, Netflix’s initial version and business model was based on physical delivery. After having spent a lot of money on the clearance of rights acquisition of movies from Hollywood studios and other filmmakers, Netflix established itself as a an audiovisual producer after realized that its initial venture would not have been successful. As a result of spending cuts on third-party films and products and investing savings on internal productions, several economic studies have shown a sharp reduction in the number of movies being simultaneously offered on Netflix. From an economic point of view, the platforms’ “silos model”, which obliges film and TV lovers to subscribe to numerous platforms, is unsuitable and very expensive especially for young viewers.
Speaker Prof. Javier FERNÁNDEZ-LASQUETTY said that films made by AI will strengthen the production forces of independent creators, because of the limited budget compared to the big majors. Perhaps the application of Artificial Intelligence will provide us with data that will allow independent producers to generate high-quality films at lower costs. However, Prof. Giuseppe MAZZIOTTI wonders whether these technologies will be accessible and broadly affordable in terms of licensing fees, assuming that independent producers will not be able to produce these technologies by themselves. Prof. Javier FERNÁNDEZ-LASQUETTY asserted that videos used to be limited to professionals, but now they are accessible to anyone with the right tools. Technological advances would allow them to generate content of special quality.
Questions from the audience
1. How can the copyright-ability of AI-generated works be measured? Can the creativity test established in the CJEU’s «Painer» and «Infopaq» decisions be applied to Irish authorship Law, to assess the human involvement requirement of these works?
Speaker Prof. Giuseppe MAZZIOTTI responded by saying that Irish legislation has faced as many problems as the United Kingdom legislation with these findings of the Court of Justice, even for much less than AI-created works. In the audiovisual sector, the European Union imposed in 1993 to include film directors as co-authors of movies. Now, if you read Irish or UK copyright statutes, you will find it strange that the authors of a film are the producer and its director. This is the result of the push coming from Europe.
Prof. MAZZIOTTI, Prof. Javier FERNÁNDEZ-LASQUETTY and Prof. Giuseppina D’AGOSTINO agreed that the best is yet to come for lawyers, in the coming years as new issues emerge.
Speaker Prof. Javier FERNÁNDEZ-LASQUETTY provided further details on the creative requirements to be met. Indeed, we can use three criteria (Indicated by AIPPI in its works about this matter) to determine whether or not there was a sufficient intervention of humans, and consequently, an author. First of all, the data selection (provided by a human or a machine?); secondly, the intervention of the human during the process of creation; and finally, the selection of the outcome. Based on the results obtained, different possibilities could be applied for nuanced situations where authorship is recognized or others where it is unclear.
2. To what extent can we tolerate the systematic use, if not too invasive, of AI in film creation or staging (use of holograms, special effects, rejuvenation of actors through new technologies)? Do you think that AI will be as essential as an actor or a director in the future?
Speaker Prof. Giuseppe MAZZIOTTI sees a clear phenomenon of convergence of different industries and types of audiovisual production in the near future. For instance, video games converge progressively towards movies, in light of their complex design and graphic, as well as the development of plots that depend entirely on choices or abilities of the players. In the same way, as in movies, video games are played interactively, raising questions about the role (and maybe rights?) of human performers of games. For instance, protection of performers’ income lies at the heart of film and TV dubbing practices in countries like France, Italy, German and Spain, where the work created by content adaptation allows them to keep a source of artistic independence. It is probably true that these fictional characters, non-human actors, and holograms, will be more widely used in the future, depending on how much appreciated they will be. The offer on the market (especially the platforms) is adapting to the plurality of their audiences.
Speaker Prof. Javier FERNÁNDEZ-LASQUETTY spoke about a matter of taste and generations, which have all developed their own art sensitivity.
3. What would happen in the case of an international jurisdictional conflict, if some AI-generated works benefit from legal recognition, from a copyright perspective?
Speaker Prof. Javier FERNÁNDEZ-LASQUETTY replied that a solution must be found under the aegis of the Berne Convention, otherwise it would be much more difficult to find a homogeneous treatment for works, especially AI creations that are highly volatile.
Speaker Prof. Giuseppe MAZZIOTTI also reaffirmed the importance of the Berne Convention and its possible, future reform to solve this problem. We shouldn’t forget the distinction between authorship and ownership. From a copyright perspective, the protection of ownership is and will be more solid and internationally accepted than authorship, even by courts, for business reasons and converging interests. It must be seen whether this type of Intellectual Property will be covered by copyright or something equivalent.
Speaker Prof. Giuseppe MAZZIOTTI emphasized the importance of being rigorous on the conclusion that a piece of audiovisual work is entirely authorless and the relationship, previously mentioned by Prof. Javier FERNÁNDEZ-LASQUETTY, between a creator and a machine-generated input, is crucial. Not to mention that we must provide a case-by-case analysis.
Finally, Moderator Prof. Giuseppina D’AGOSTINO, wondered if there were any conclusions to be drawn about biases in the data, which are a major concern in North America. A simple example could be Netflix’s suggestions: on the one hand, it remains a useful service for clients, but on the other hand it represents undercurrent algorithms that could lead to biased data.
4. What should we do to ensure more equality on data in the main platforms that we are dealing with?
Speaker Prof. Giuseppe MAZZIOTTI reminds us that the commercial audiovisual offering ultimately educates and influences audiences in terms of taste, which is an important (if not the most important) business issue. Because of the exponential volume of data and, consequently, the competitive advantage granted to the content producers (who are sometimes also the owners of platforms), the ability to control data is becoming an alarming problem. Furthermore, it is important to remember that data is a sensitive business element, in terms of trade secrecy as well. It constitutes the most important asset in this high-tech world. We are observing, at least in the European Union, an attempt to make these data-rich companies more accountable. In this regard, the EU is currently implementing two significant reforms: the Digital Market Act and Digital Services Act. The approval and entry into force of The Digital Marketplace Act might be more relevant in the audiovisual sector because it aims to help third-party companies that thrive on a platform have some access to data concerning their products. To North America’s dismay, the European Union could become the world leader in advocating and enforcing these data access rights, whose secrecy would otherwise strengthen an oligopolistic market, giving rise to a cultural scenario Europe clearly wants to avoid.
Speaker Prof. Javier FERNÁNDEZ-LASQUETTY agreed with this point of view. He also pointed out that not only the GDPR but also Competition Law and consumer protection will force platforms to comply with certain requirements. The question of data, economy and AI-generated data represents an issue for European authorities. Indeed, from one side, the balance between sharing and keeping the information should be treated under the Competition Law. On the other side, there are fewer concerns to have, because the influence of the platform will encourage to consume as many products as possible. Therefore, it would not always push towards thrillers or comedies but maybe towards something else, in one way or another.
As a conclusive remark, Prof. Manuel Desantes Real reaffirmed the importance of this topic, which questions everything: Who is the author? Who is the owner? When and where will Artificial Intelligence arrive? Technology will certainly allow people to produce audiovisual works for a very limited cost in the near future, whether the creator is human or not. He concludes that this debate also embraces broader and practical aspects, including privacy, liability, creativity, and the future of what culture means.
Report by Adele Luisa Serio and Xiao Baiyang