The Secretariat for the Economy and the Holy See’s Labour Office organize an event on the potential and challenges of artificial intelligence, with the 'appreciation and encouragement' of the Pope.
By Edoardo Giribaldi “An abundance of means and a confusion of ends.” This phrase, attributed to Albert Einstein, offers a snapshot of a world challenged and shaped by new technologies. The interests at stake are multiple and not “neutral.” In this context, the Holy See — which has no military or commercial objectives — can play a key role in promoting global governance capable of developing systems that are “ethical from their design stage.” These were some of the themes highlighted during the seminar “Potential and Challenges of Artificial Intelligence,” organized today, Monday 2 March, in Rome, at the Salone San Pio X on Via della Conciliazione 5, by the Secretariat for the Economy and the Office of Labor of the Apostolic See (ULSA). The proceedings were opened by Professor Pasquale Passalacqua, Director of ULSA, who noted that Pope Leo XIV himself, informed of the initiative by its president, Monsignor Marco Sprizzi, had “appreciated and encouraged” it, expressing the hope for “deeper awareness in this highly relevant and complex field.” The discussion was moderated by Alessandro Gisotti, Deputy Editorial Director of the Dicastery for Communication. Speakers included Bishop Paul Tighe, Secretary of the Dicastery for Culture and Education; the Franciscan friar Paolo Benanti, professor at the Pontifical Gregorian University and at Luiss Guido Carli University; and Professor Corrado Giustozzi, lecturer in the Master’s Degree program in Intelligent Systems Engineering at Rome’s Campus Bio-Medico University. To summarize the consequences of the widespread uptake in 2022 of ChatGPT, Bishop Tighe used the acronym VUCA: Volatility, Uncertainty, Complexity, and Ambiguity. In this regard, he referred to recent developments and the case of Anthropic, a U.S.-based company founded with the aim of promoting more ethical AI and reportedly subject to “government pressure to relax its ethical commitments regarding military and surveillance uses.” The development of new technologies, he noted, does not occur in neutral spaces but is intertwined with “geopolitical rivalries, commercial pressures, and personal ambitions.” Faced with such complexities, the Bishop referred to the document Antiqua et nova, which points to the “wisdom of the heart, capable of integrating the whole and its parts,” as what humanity most needs today. For its part, the Church possesses “moral authority” and the ability to bring together qualified interlocutors, thereby becoming a meaningful partner in guiding the development of AI. Drawing together the themes of the intervention, Gisotti underlined that the seminar also represented a “commitment” on the part of the ecclesial community. Father Benanti’s presentation focused on the ethical challenges of artificial intelligence, proposing a new “ethics of technology” that questions the “politics” embedded in such models. “Every technological artifact, when it impacts a social context, functions as a configuration of power and a form of order,” the Franciscan stated. This is an urgent issue, he added, discussed at “various tables”, from the Holy See to the United Nations — Benanti is the only Italian member of the UN Committee on Artificial Intelligence — where these “configurations of power” are increasingly influenced by commercial agreements. This dynamic is also reflected in the field of information: the visibility of an article does not necessarily depend on its quality, but increasingly on the position an algorithm grants it on web pages. It is a “mediation of power,” Benanti concluded. Professor Giustozzi’s intervention focused on the nature of one of AI’s essential components: the algorithm, and the critical issues associated with decision-making processes based on it. Among these is the problem of bias: algorithms may at times incorporate prejudices — whether unintentional or deliberate — that distort results or render them inequitable. In this sense, training — the phase in which an algorithm is developed through data input — assumes decisive importance: if the data is incomplete or distorted, the results will inevitably be erroneous or discriminatory.