CONSUMER PROFILING IN AIR TRANSPORT AND TOURISM that can only be deciphered by an expert32. Therefore, the Authority is not very well able to know how the systems work, nor to assess whether the determination of prices is the expression of a calculation functional to altering the ordinary dynamics of the market. In essence, it is no longer possible to establish whether the distorting effect results in an agreement restricting competition or in the abusive exercise of a dominant position, because the investigation stops beforehand, i. e. when faced with the lack of a full and certain analysis of the decision - making process; if the operation of the algorithm cannot be guessed at, it cannot be demonstrated that it was constructed for illegitimate purposes. Furthermore, algorithms may create errors, which could spoil the final evaluation, despite the fact that the relevant variables are apparently neutral; in particular, they "could use" an element "other than a discriminating factor"33, but, insofar as it is "statistically correlated with the latter", produce then “a similar discriminating effect, at the price, however, of being more difficult to perceive even by those who programme and supervise the operation"34. These difficulties in application raise a more general question, namely whether objective and more drastic regulatory techniques could not guarantee better protection. Article 1 of Decree - Law No. 104 of 2023 in the wording prior to its conversion into law had its grounds, since it prohibited the dynamic fixing of tariffs if this resulted in a price two hundred per cent higher than the average price of the flight; there was a limit anchored to an economic criterion, free from discretionary evaluations, based on a simple finding. What is different is the case of profiling, which, for instance, provides personalised advertising or leads to different commercial offers depending on the nature of the person or the category in which he or she is included; 32 Cf. P. Zellini, La dittatura del calcolo, Milan, 2018, 16 ff.; C. Accotto, Il mondo dato. Five short lessons in digital philosophy, Milan, 2017, 64 ff.; C. O'Neill, Weapons of mathematical destruction. How big data increases inequality and threatens democracy, Florence, 2017, 34 ff.; V. Mayer - Schönberger - K. Cukier, Big data. A revolution that will transform the way we live and already threatens our freedom, Milan, 2013, 52 ff. 33 Cf. M. V. Ballestrero, Ancora sui raiders. La cecità discriminatoria della piattaforma, in Labor, 19 January 2021; M. Barbera, Discriminazioni algoritmiche e forme di discriminazioni, in Lab. law issues, 2021, 1, 20 ff.; A. Perulli, La discriminazione algoritmica: brevi note introduttive a margine dell'ordinanza del Tribunale di Bologna, in Lav. dir. Europa, 2021, 1 ss.; M. Borzaga, Qualche riflessione a partire dai primi orientamenti della giurisprudenza di merito in tema di discriminazioni algoritmiche, in Ianus - dir. fin., 2021, 24, 151 ss. 34 See G. Gaudio, Le discriminazioni algoritmiche, loc. cit., 6 ff.; A. Lo Faro, Algorithmic decision making and management of labour relations: cosa abbiamo imparato dalle piattaforme, in Federalismi.it, 5 October 2022, 195 ss.; S. Renzi, Decisioni automatizzate e analisi predittive e tutela della privacy dei lavoratori, in Lav. dir., 2022, 593 ss.
RkJQdWJsaXNoZXIy MTE4NzM5Nw==