CONSUMER PROFILING IN AIR TRANSPORT AND TOURISM pays attention to the way in which fees and their individual components are disclosed, given their impact on the overall transparency of prices and their comparability; in fact, travellers must be aware of the changes that have determined the fees27. At least in this respect, the survey is very ambitious, because it presupposes that companies disclose their calculation mechanisms. Moreover, albeit in the different sphere of labour law28, the articulation of such systems has been highlighted and it has been emphasized that, with machine learning, “the possibility of obtaining a possible ex post explanation of the decision requires recourse to reverse engineering methods that have significant costs and, above all, grow as the complexity of the algorithm increases. The result obtained from these analyses would be (...) comprehensible, in most cases, only by an expert subject", and "a subsequent intervention would then be necessary to facilitate comprehension by a non - expert subject, such as (...) the employer himself"29. Recently, the Authority was asked about the existence of a restrictive agreement with regard to the economic conditions demanded by some so - called low cost companies on routes at the ends of which were, on the one hand, Palermo or Catania and, on the other, Rome, Milan, Turin or Bologna30. In particular, with reference to some carriers, the substantial increase in fees that occurred over the Christmas period of 2022 was allegedly due to a collusive intent. The Authority observed how companies define their proposals on the basis of decisions of a statistical nature and how well they calculate fares in relation to available seats depends on the balancing of interests, i. e. the maximum increase in average revenue and the need to fill the aircraft. 27 Cf. C. Romeo, L'era degli algoritmi e la sua incidenza nell'ambito della certezza del diritto: un connubio sospetto, in Lav. giur., 2024, 6 ff., according to which “the algorithm cannot replace (...) the human mind, in the perception that algorithmic use in the choices to be profiled must ensure efficiency and speed, but also impartiality of the choices reserved to the operator, capable at any time of intervening with the necessary spaces of discretion”; P. De Hert - H. Lammerant, Predictive profiling and its legal limits: effectiveness gone forever?, loc. cit., 145 ff. 28 Cf. M. Peruzzi, Artificial Intelligence and Work. Uno studio su poteri datoriali e tecniche di tutela, Turin, 2023, 13 ff.; S. Watcher - B. Mittelstadt - C. Russell, Counterfactual explanations without opening the black box: automated decisions and GDPR, in Harv. journal law tech., 2018, 841 ff.; G. Gaudio, Algorithmic management, employers' powers and burden of proof: in search of the material truth behind the algorithm, in Lab. law issues, 2020, 27 ff. 29 Cf. G. Gaudio, Le discriminazioni algoritmiche, loc. cit., 6 ff. 30 See Order No. 31014 of 2024, published in the Competition and Market Authority Bulletin No. 1 of 2 January 2024.
RkJQdWJsaXNoZXIy MTE4NzM5Nw==