Nouveautés

La justice prédictive représente un outil permettant à un professionnel du droit de quantifier statistiquement ses chances de succès en cas de procès. Il permet également une analyse sémantique des décisions de justice, en fonction du vocabulaire employé par le juge. Ces outils prédictifs sont basés sur le machine learning. Ce sont des algorithmes autonomes, qui à partir des données qu’on leur fournit, vont pouvoir reproduire une tâche plus rapidement que l’être humain. Cette technique qui relève de l’intelligence artificielle a notamment été développée pour la reconnaissance d’images. Aujourd'hui, des acteurs privés, les legaltechs, développent ces algorithmes pour les appliquer au monde du droit. L’objet de ce travail de recherche est d’étudier l’impact de cette évolution technologique sur la pratique du juge, de l’avocat et aussi sur celle du juriste d’entreprise. Le contentieux de la propriété intellectuelle a été utilisé comme un exemple d’appui pour déterminer quelles seraient les difficultés d’application de la justice prédictive aux contentieux plus complexes. Il est à noter qu’à ce jour, la justice prédictive ne s’applique qu’aux contentieux répétitifs comme les demandes d’indemnisation suite à un licenciement ou le calcul de prestation compensatoires en cas de divorce. Il s’agit donc d’un travail prospectif.

La marque étant un « facteur de ralliement du public, en ce sens qu’elle représente le lien entre le produit ou service et le consommateur », elle se doit d’avoir une certaine force publicitaire et promotionnelle. Devant attirer en elle-même la clientèle et l’inciter à porter son choix sur tel produit ou service plutôt que sur tel autre, la seule distinctivité du signe ne suffit plus : l’image de marque doit renvoyer à un univers à la fois commercial et culturel, voire artistique. Pour ce faire, existe-t-il pratique plus idoine que le dépôt d’une marque constituée par une création de forme protégée ? Plus particulièrement, la notoriété vertigineuse dont bénéficient les personnages fictifs permet de procurer un solide caractère attractif à la marque. Nous entendons par l’appellation « personnage fictif » ou « personnage de fiction » toute création revêtant une forme humanisée, animale ou purement fantaisiste, à l’exclusion des individus ayant réellement existé. Il est remarquable de constater que les marques constituées par des personnages fictifs sont beaucoup plus caractéristiques que la plupart des signes habituellement choisis. Sous l’effet d’un character merchandising hérité des Etats-Unis, il est commun de voir des personnages se détacher de leur oeuvre originelle et se décliner en une multitude de produits dérivés. Aussi, les personnages font ils l’objet d’une lutte commerciale et d’une course au dépôt, opposant les différents exploitants commerciaux du marché mondial. Plus encore, « le succès immense de toutes les opérations commerciales lancées jusqu’à présent (qu’il s’agisse par exemple de Snoopy, Mickey ou Tarzan) démontre que plus le personnage devient envahissant, plus le public l’apprécie ». Il s’agira ainsi de créer des marques nominales, figuratives ou semi-figuratives, que le personnage ait acquis son identité dans une oeuvre littéraire ou artistique qu’il ait été créé spécialement pour la commercialisation d’un produit donné. Sur ce dernier point, nous pouvons citer, parmi tant d’autres, le célèbre Bonhomme Michelin créé en 1898 dans le domaine automobile. En 2000, il fût élu meilleur symbole de marque de tous les temps par un jury international réuni par le Financial Times. Bien que faisant traditionnellement partie d’un décor initial, à l’image du roman, du film ou de la bande dessinée, le statut juridique autonome des personnages fictifs fût très tôt reconnu par les tribunaux sur le terrain du droit d’auteur. L’acquisition immédiate du droit d’auteur sur la création, contrairement au droit de marque qui suppose un dépôt, implique nécessairement que la première strate est celle du droit d’auteur, la seconde, éventuelle, est constituée par la marque. Tandis que le dépôt d’une marque constituée par un personnage protégé par le droit d’auteur est hautement pratiqué - dans la mesure où l’empilement de droits sur un même objet conduit à conférer à l’ensemble un rayonnement plus étendu et une protection solidifiée-, il n’existe aucun corps de règles dans le droit positif français, qu’il s’agisse de la loi ou de la jurisprudence, pour régir une telle situation. Les différents modes d’influence du droit d’auteur sur la validité de la marque de personnage seront étudiés successivement, du plus positif ou plus prospectif, du plus consensuel au plus discuté. --------------------------------------------------------------------------------------------------------------- Since the brand is a rallying factor for the public, in that it represents the link between the product or service and the consumer, it must have some advertising and promotional force. In order to attract customers and encourage them to choose one product or service (rather than another), the brand image must refer to both a commercial and cultural -even artistic- environment. To achieve this, is there a more convenient practice than the filing of a trademark constituted by a creation of protected form? More particularly, the tremendous fame enjoyed by fictional characters provides a strong attractive character to the brand. By the term "fictional character" we mean any creation that takes on a humanized, animal or purely fanciful form (exclusive of individuals who actually existed). It is remarkable to note that trademarks made up of fictional characters are much more characteristic than most of the signs usually chosen. Under the effect of a character merchandising inherited from the United States, it is common to see characters detach themselves from their original work and decline into a multitude of derivative products. In this regard, characters are subject to a commercial fight between the different commercial operators of the world market to file trademark applications. Moreover, the great success of all the commercial operations launched so far (whether Snoopy, Mickey or Tarzan, for example) demonstrates that the more invasive the character becomes, the more the public appreciates them. Either the character has acquired his identity in a literary or artistic work, or it has been created especially for the marketing of a given product. On this last point, we can mention, among many others, the famous Michelin man created in 1898 in the automotive field. In 2000, he was voted best brand symbol of all time by an international jury assembled by the Financial Times. Although traditionally part of an initial setting, like the novel, the film or the comic strip, the autonomous legal status of the fictional characters was very early recognized by the courts on the ground of copyright. The immediate acquisition of a copyright on the creation -unlike trademark rights which presuppose a filling- implies that the first stratum is that of the copyright, the second, if any, is constituted by the trademark. The filing of a trademark made of a character already protected by copyright is highly exercised, to the extent that the accumulation of rights on the same object leads to a solidified protection. However, there isn’t a set of rules in French law to govern such a situation. The different influences of copyright on the validity of the trademark character will be studied successively, from the most positive to the most prospective, from the most consensual to the most discussed of it.

Corruption is a conspicuous phenomenon in developing countries like China and India. Indian economist Basu proposed to legalize bribe-giving in 2011, which have given rise to great and heated debates between the economists and socialists on Basu’s proposal in recent years. Basu pointed out that in countries like India, there is a common situation when the bureaucrat often demands some “extra fee” to provide a public service (delivering a driver's license, a visa…) which is supposed to belong to the citizen. In such a situation, which is called “harassment bribe” by Basu, an honest citizen should be less punished by the law and should be given leniency if they report the bribery. The economic theoretical analyses on bribery is either by microeconomic method or by models. The latter often includes principal-agent model and game theory. Game theoretical analysis is most frequently used to analyze Basu’s proposal as it can reflect the strategic interactions of the individuals in a corrupt behavior. Even though the individuals are assumed to be purely rational, the social, cultural and psychological factors sometimes can have more significant effects on individuals’ behaviors. That’s why the experimental studies based on a bribery game is not always in accordance with the game theoretical analyses. The literature about encouraging the whistle-blowing in the corrupting behavior can be traced back to Rose-Ackerman (1975, 1978). And one of the earliest experimental analyses based on bribery game was realized by Abbink et al. in 1999, which affirmed the three important characteristics of bribery are reciprocity, negative externalities and its inherent risk. To break the trust between the two corrupting parts, Spagnolo et al. (2005) firstly compared corruption to cartel relationship, and they deemed that asymmetric penalties and leniency may be effective in breaking the inner trust in the illegal transaction and then can be used to deter the corruption. Nell (2007) then built a bribery game in adopting different penalties on the bribers and the bribees before and after the reciprocity. And he found that a high penalty on bureaucrat for doing the favor and a low penalty on briber for giving a bribe will deter the bribery effectively. However, by reviewing his game true and hypotheses, I find that generally, a low fine on briber, a low fine on bureaucrat before the reciprocity while a higher one afterwards will effectively encourage either part to whistle the blowing or cut off their benefits of being corrupt. Furthermore, a total leniency on briber is also effective on elimination of the corruption trade but and only if the briber reports the bribery after the reciprocity. However, Abbink et al. in 2002 confirmed that encouraging the whistle-blowing in a collusive bribery (“non-harassment bribe”) is not effective in anti-corruption, especially when the game is repeated (Spagnolo et al.in 2015). Some dishonest citizens may also abuse the policy and frame up the honest bureaucrat. This concern is dismissed by Serra, who brings up the idea of a combination of “bottom-up” and a “top-down” mechanism to guarantee the justice of the legislation. As Nell completed, the low penalty on bribers will backfire when the dishonest bribers may use the leniency policy to threaten an honest bureaucrat to get what he is not entitled to. Basu’s proposal seems to resolve the backfire worries of Nell if we can successfully pick out the “harassment bribe” scenarios, which is not easy especially when the criteria are not objective. For example, using the bribery game tree and the hypotheses of Spagnolo in 2012, I notice that if fulfilling all the assumptions, the “harassment bribe” happens only when the cost of delivering the license for the bureaucrat is positive. Besides, Basu’s proposal (which in this case equals a double penalty on the bureaucrat and the bribe is refunded to the briber who blows the whistle) only works when the cost of whistle-blowing (pecuniary, mental, or moral) is less than the bribe amount, and leniency policy works if the penalty on the bureaucrat is big enough (Verma and Sengupta in 2015). Even though we can successfully identify a “harassment bribe”, and the cost of reporting is trivial, a bribe amount may become higher and accordingly, bribery may transfer to higher level of officials with Basu’s proposal (Oak in 2013, Abbink et al. in 2013, Verma and Sengupta in 2015). The bureaucrat who is reported may get rid of the conviction of the corrupt crime, or they can retaliate the citizens once they get back to their job, which is often the case because the “harassment bribes” are often petty corruption and the bureaucrat often can keep their official position (Abbink et al. in 2013 and 2014, Apesteguia et al. in 2007, Basu in 2016, Bigoni et al. in 2015, Engel et al. in 2012, Oak in 20113, Hinloopen and Soetevent in 2008, Spagnolo et al. in 2017, Verma and Sengupta in 2015, …). What’s worse is that in countries where the corruption is conspicuous like China and India, the bureaucrat often has such an arbitrary power that they can even change a “harassment bribe” scenario to a “non-harassment bribe” scenario (Oak in 2013, Li, X 2012). That’s why the statistics in corruption after the reform in 1979 in China shows no sign of effectiveness and success of the anti-corruption policy with asymmetric punishment and leniency (Spagnolo et al. in 2017). The mechanisms such as rotation of staff between the officials (Abbink 2012), the “bottom-up” and a “top-down” mechanism (Barr and Serra 2009), the competition between the officials (Ryvkin & Serra in 2013), a platform such as http://ipaidabribe.com/ (Serra in 2016) to encourage the whistle-blowing may help weaken the monopoly power of the bureaucrat, but the effects of these external factors are always limited. An experiment by comparing the effectiveness of Basu’s proposal and leniency found that more bribery cases appear under asymmetric punishment in China compared with Germany (Abbink et al. in 2013). Kingston (2007) have already pointed out that the social structure can impact the corruption. By building a parochial corruption game, he confirmed that in a society in lack of spirit of contract and confidence, the social network and connection will “kidnap” the individuals to take participate into corruption for the collective interests of the social group which they belong to, especially in the long term.