“This article was automatically written by Tencent Dreamwriter robot”

“This article was automatically written by Tencent Dreamwriter robot”

Nanshan’s Court got certainly worldwide noticed for a pivotal decision, namely “Shenzhen Tencent Computer System Co., Ltd. (“Tencent”) vs Shanghai Yingxu Technology Co., Ltd. (“Yingxu Technology”)” (full text available here).

With this judgement, the Chinese Court ruled in favor of recognizing protection to AI-generated content under Copyright Law.  

Specifically:

  • In August 2018, Tencent published an article on the Shangai stock exchange index on its website, which terminated with the following statement: “This article was automatically written by Tencent Dreamwriter robot”;
  • The article was written with the support of the software Dreamwriter;
  • Tencent claimed that Dreamwriter computer software was a data set and algorithm-based intelligent writing assistance system, independently developed by Tencent Technology (Beijing) Co., Ltd (an affiliate of Tencent), then licensed to Tencent (see claim here);
  • Yingxun Technology published the same article on the Shangai stock exchange index on its website;
  • Tencent claimed, inter alia, copyright infringement since, according to its point of view, the article on Shangai stock exchange was protectable under copyright law and the rights were attributable to it;
  • The Chinese Court (specifically, the Court of Nanshan) ruled in favor of recognizing protection to an article written with the support of an AI under Chinese Copyright Law, sustaining – in brief – that the article (a) could be included within the scope of protection being a literary work and that (b) “creative choices” were made by the team who selected the data to be included in the AI which then was involved in the making of the work.

The decision here addressed represents a pivotal decision with reference to AI-generated content protection.

Notwithstanding the above, please note that of course the discussion on protectability and ownership of AI-generated content is still open at a worldwide level.

Francesca Di Lazzaro and Maria Di Gravio

Court of Nanshan (District of Shenzhen) 24 December 2019 – Case No. (2019) Yue 0305 Min Chu No. 14010, Shenzen Tencent Computer System Co., Ltd. vs Shanghai Yingxu Technology Co., Ltd.

Copyright protection of algorithms does not prevent the disclosure of their source code in the context of administrative proceedings

Algorithms are often used for managing complex administrative proceedings where multiple data and parameters have to be analysed to produce a result. Since algorithms can be protected under copyright laws as software (including their source code), it is questionable whether copyright protection might limit the right to access of interested parties in administrative proceedings. In two recent cases (here and here), the Italian Administrative Court of Lazio (TAR Lazio) has clarified the nature of the electronic administrative document and the scope of the right to access pursuant to Law n. 241/1990 with regard to the source code of an algorithm compiled by a software house on request by the Public Administration. The cases at stake have been promoted by a number of Italian trade unions against the Ministry of University and Education (“MUIR”) with the purposes of gaining access to the source code of the algorithm used by MUIR to manage the territorial relocation of school professors under mobility procedures.

Upon first request, the MUIR refused access to the source code of the algorithm developed by a software house on MUIR’s request on basis of the following arguments: (i) the source code itself cannot be considered part of the electronic administrative document and, consequently, does not imply the right to access of interested parties in administrative proceedings, and (ii) the source code enjoys the copyright protection as software and the access to the source code would prejudice the intellectual property rights of the software house. More in detail, MUIR has held that the disclosure of a document describing the way of functioning of the algorithm could be considered sufficient protection for the trade unions and that the Legislative Decree n. 97/2016 (Art. 6) on the civic right to access (for preventing corruption and enhancing transparency in the public sector) expressly excludes access to the acts of the Public Administration when the access could prejudice the economic interest of private parties, thus included their intellectual property rights.

In the Administrative Court’s opinion, the MUIR must allow access to the source code of the algorithm since it can be considered part of the administrative proceeding subject to the right to access of interested parties. MUIR has requested the software house to compile the algorithm with the specific purpose of managing in electronic form the public procedure of territorial relocation of school professors under mobility, according to public rules and collective employment agreements. From a structural point of view, the outputs of the algorithm: (i) are the results of the combination/elaboration of data collected in various endoprocedural acts and (ii) make application of the public rules on territorial mobility.

Taking into consideration the ratio of the right to access in administrative procedures, also the source code of the algorithm enjoys the nature of electronic administrative document and such nature implies that right to access should be allowed also with regards to algorithm. Reasoning to the contrary will lead to the unacceptable consequense that the right to access could be automatically excluded by decision of the Public Administration to manage the administrative proceeding by electronic means. TAR Lazio further clarified the notion of electronic administrative document which, in the Court’s opinion, should not include only those administrative documents formed via electronic means (for the purpose of documentation) but should also include those administrative documents where the elaboration of contents and data (for the purpose of issuing an output) are taken into account.

Also the copyright protection of software (which encompasses also the source code) has not been considered by the Court as an argument for excluding the right to access to the algorithm. First of all, TAR Lazio acknowledges that software can be protected under copyright laws not only as an informatic language but also as a creative work resulting from the use of a certain informatic language. In the case at stake, the algorithm is a software created for a specific purpose of the Public Administration and, in the absence of any indication to the contrary in the agreement between the PA and software house, can be assumed that the software house has transferred to the PA all the economic rights in the algorithm. In the Court’s opinion, the nature of creative work of the algorithm should not interfere with the right to access in the administrative proceedings of interested parties, since the right to access does not prejudice the right to exploitation of intellectual properties (any reproduction made by the interested parties is functional to the exercise of rights to control the administrative proceeding only and not to the commercial exploitation of the algorithm).

In addition, TAR LAZIO considered that is not relevant for excluding the right to access to the source code of the algorithm the fact that: (i) the source code is a pure informatic language unreadable by the public officers and written by a private company (i.e. the software house on behalf of the PA) and (ii) the source code is compiled for the mere application of public rules and collective labour agreements, which are accessible themselves even without direct access to the source code. The Court ruled in favour of the right to access to the source code also on the basis that what impact the giuridical position of private individuals are the outputs of the algorithm.

These interesting administrative rulings offer a clear and deep reconstruction of the notion of electronic administrative document (expanding such notion to include also algorithms) but should be subject to further analysis with regards to the asserted strike of balance between the right to access and the protection under copyright laws of the source code, exspecially taking into consideration possible future cases where the PA should make use of algorithms: (a) not specifically developed for a single administrative proceeding (under the assumption of a complete transfer of intellectual property rights) and/or (b) based on more sophisticated technologies licensed to the PA under a proprietary scheme.

Gianluca Campus

TAR Lazio, case No.  3742/2017, 21 March 2017, CISL, UIL, SNALS Vs MUIR (President of the Court: Hon. R. Savoia; Judge-Rapporteur Hon. M.C. Quiligotti)

TAR Lazio, case No.  3769/2017, 22 March 2017, Gilda Vs MUIR (President of the Court: Hon. R. Savoia; Judge-Rapporteur Hon. M.C. Quiligotti)

Italian Courts and the protection of Graphical User Interfaces under copyright laws: some recent cases

In some recent cases (case “Trend for Trend Vs Pix Agency” dated 5.2.2016; case “DriveK Italia Vs Webbdone” dated 20.3.2016 and case “Esko Software Vs Beegraphic” dated 6.7.2012) the Court of Milan has ruled with regard to the legal protection of the Graphical User Interfaces (“GUI”) and more in detail on the protection of the overall aspect of the GUI (the so called “Look&Feel”). The GUI can be considered as an atypical and complex creative work based on a project that combines graphical and navigation elements allowing interaction between users and devices (both at software and hardware level).

The term “Look&Feel” is sometimes used in EU case law (AG Bot – C 309/09) and by Italian scholars (Giordano D., Software e grafica dei social network, AIDA 2011, 248) as a synonymous of GUI (see AG’s Conclusion – C 309/09: “The graphic user interface, commonly referred to as the ‘look and feel’, enables communication between the program and the user”). More precisely, some scholars (Hayes, What’s left of “look and feel”: a current analysis, in Comp. Lawyer, X 1993, n. 5) consider “Look” the combination of all the visual and/or audio elements of a computer program as presented in a GUI and “Feel” as the sequence, structure and organization of the non-literal elements into a GUI. It has been questioned whether the overall “look and feel” (i.e. the combination of visual elements and the organization of such elements for interaction with users) can be protected as a whole.

The overall “Look&Feel” and the individual elements in a GUI can be protected, under certain conditions (notably with an adequate level of creativity), by Italian copyright law (see Article 1.1 and Article 19 Italian Copyright Law). This conclusion should not be disputed in case of one-to-one copies of GUIs and/or individual elements in a GUI (i.e. in case of “literal copyright infringements”). More problematic is the case of non-literal copyright infringement of the overall “look and feel”, where it has to be clarified which elements in the overall “look and feel” are eligible for copyright protection.

An obstacle to the protection of the “Look&Feel” is represented by Article 1.2 Software Directive (2009/24/EC), according to which ideas and principles of a computer programs are not eligible for protection under copyright law (“Protection in accordance with this Directive shall apply to the expression in any form of a computer program. Ideas and principles which underlie any element of a computer program, including those which underlie its interfaces, are not protected by copyright under this Directive”). Such EU provision is implemented by Article 2 n.8 Italian Copyright Law. It could be argued that “Look&Feel” incorporates some basic ideas and principles of the computer program generating the GUI. According to the above, protection under Italian copyright law could be granted to the overall “Look&Feel ” without taking into consideration those elements in the GUI: (i) that are expression of technical ideas and principles at the basis of the computer program generating the GUI; and (ii) that are the necessary result of the functional organization of the GUI.

On the other hand, the “Look&Feel” of a GUI could imply a level of creativity in the sequence, structure and organization of the elements, eligible for copyright protection as “elaborated ideas”. A possibile confirmation of this type of protection can be found in a EU case related to the protection of a computer manual. The CJEU has stated that, with reference to a computer manual, also the choice, sequence and combinations of commands, options, defaults and iterations is expression of creativity and can result in an intellectual creation protected under copyright law (see CJEU case C-406/10 “[…] the keywords, syntax, commands and combinations of commands, options, defaults and iterations consist of words, figures or mathematical concepts which, considered in isolation, are not, as such, an intellectual creation of the author of the computer program. It is only through the choice, sequence and combination of those words, figures or mathematical concepts that the author may express his creativity in an original manner and achieve a result, namely the user manual for the computer program, which is an intellectual creation”).

Italian case law regarding the alleged copyright infringement of the GUI of a website has usually excluded the violation of copyright laws in case of non-literal infringement and has made in some cases application of the rules against unfair competition practice in the Italian Civil Code (namely article 2598 ICC). Nevertheless in their legal analysis the Courts have expressed their view on the protection of GUIs under copyright laws. The Court of Milan in the case Trend for Trend Vs Pix Agency dated 5.2.2016, regarding the alleged copyright infringement of the GUI of a website, held that the scope of copyright protection includes both the expression and the elaboration of ideas but found that the format of the allegedly infringed website (inclusive of a description of the GUI) filed with the SIAE was not eligible for protection and that a comparison of the GUI (without taking into consideration forms imposed by the technical nature of the services proposed by the websites) excluded any graphical coherence between the GUIs. In another case the Court of Milan dated 20.3.2016 (DriveK Italia Vs Webbdone), excluding copyright infringement of the GUI of a computer program, stated that it is on the claimant to prove the level of creativity of the GUI, also by offering a reconstruction of the state of the art in the sector of the software at stake and/or by offering a comparison with other comparable software. Finally, in a case before the Court of Milan dated 6.7.2012 (Esko Software Vs Beegraphic) the technical expert found that most of the elements of the GUIs under comparison were imposed by technical reasons and that the organization of the elements in the GUI of the defendant was sufficiently differentiated by the same elements in the GUI of the claimant.

Gianluca Campus

Oracle v Google: fair use defense saved web developers!

On 26 May 2016 a jury in the District Court for the Northern District of California unanimously upheld that Google’s use of Oracle’s APIs constitutes a “fair use” under the U.S. Copyright Act (full text here). The battle between Oracle and Google begun in 2010 when Oracle sued Google for using JAVA APIs, owned by Oracle, without permission. API stands for Application Programming Interface and is a set of instructions which allows different types of software to communicate to each other. When Google started the implementation of its Android Operating System (OS), despite writing its own version of JAVA, it copied some features of Oracle’s APIs in order to facilitate app developers, already familiar with them, in writing programs for Android. Therefore Oracle claimed a copyright on JAVA APIs  and the infringement of the copyright by Google.

On 31 May 2012 Judge Alsup (Northern District of California) ruled that APIs cannot be subject to copyright under the U.S. Copyright Act (full text here) which does not protect purely functional things, i.e. “process, system or methods of operation” (17 U.S. Code § 102). He motivated that in the situation at issue the taxonomy “is composed entirely of a system of commands to carry out specified computer functions”, therefore, “to accept Oracle’s claim would be to allow anyone to copyright one version of code to carry out a system of commands and thereby bar all others from writing their own different versions to carry out all or part of the same commands”. Oracle appealed and on 9 May 2014 the U.S. Court of Appeals for the Federal Circuit reversed the ruling (full text here), establishing that JAVA APIs are copyrightable, while leaving open the possibility of a “fair use” defense in Google’s favor. In reaching its conclusion, the Federal Circuit underlined that “a set of commands to instruct a computer to carry out desired operations may contain expression that is eligible for copyright protection”. Therefore, even if a computer program is functional by definition, according to the Court, can be protected by copyright as long as it is original and its underlying idea could have been expressed in multiple ways by the author. On October 2014, Google filed a petition to the U.S. Supreme Court requesting the review of the Federal Circuit’s decision. The U.S. Supreme Court denied Google’s petition on June 2015. Therefore, the case returned to the District Court and a jury agreed that Google’s re-implementation of JAVA APIs is protected by fair use.

This decision constitutes a big relief for all developers out there which are used to re-implement each other’s APIs in performing their everyday work. A different finding by the jury could have caused a chaos in the “programmer’s world” as it could have been the begin of numerous copyright claims with a consequent slowdown of the IT progress, at least in the long run. Nevertheless, the most founded decision was Judge Alsup’s ruling in 2012, when he ruled that computer programs cannot be protected by copyright law. In fact, all programs are functional for their own nature, therefore the consideration that the same functional scope can be performed by a variety of software is without relevance in order to establish the copyrightability of computer programs. What matters from a juridical standpoint is that in computer programs there is not an actual (but only a fictious) dichotomy between idea and expression (which is the core requirement for copyrightability), instead the two merge with the consequence that APIs should not be protected by copyright, as Judge Alsup’s ruled in 2012. Nevertheless, Oracle plans to appeal, let’s see what’s next!

Elisabetta Coronel Vera

(IPlens’ guest)

District Court for the Northern District of California, 26 May 2016, Oracle America, Inc. v. Google, Inc.