Conference Program

The AI Act, the European Regulation on Artificial Intelligence, (European Parliament, 2024), and the Guidelines for the introduction of Artificial Intelligence in Educational Institutions (MIM, 2025), published by the Ministry of Education and Merit, in August 2025, warn of the possible risks caused by AI with respect to our health, individual and collective safety and the protection of fundamental rights.

In both documents, reference is made to the biases, biases and biases that can occur «when AI integrates discriminatory attitudes towards one or more groups of people into its algorithms, in particular related to sex, race, religion, political opinions or particular personal and social conditions» (MIM, 2025, p. 10).

In childhood, biases represent a truly harmful risk: they are not just «errors of thought», but treacherous barriers to development. When a child grows up in prejudice, it crystallizes into mental patterns that condition his identity and his way of interacting with the world. AI trained on biased data can amplify stereotypes of gender, ethnicity, and socioeconomic status; The scenario becomes even more catastrophic when it is the algorithm itself that generates distorted and potentially harmful and dangerous outputs. The risks for children are even higher than for adults: in fact, in The Society of Transparency, the Korean philosopher Byung-Chul Han states that «everyone voluntarily surrenders himself to the panoptic gaze. One intentionally collaborates with the digital panoptic, revealing and exposing oneself» (Han 2014, p. 83). Bias thus becomes amplified discrimination due to distorted datasets that allow AI to produce social injustices (Floridi, 2022; 2025).

It is necessary to train educators and teachers so that they recognize all possible algorithmic biases: from gender bias to cultural biases, from linguistic biases to representation biases, from historical-geographical bias (Chen, 2023), which undermine the principle of justice, causing ethical, moral, social and economic damage. AI is not a neutral tool, as it reflects the values of those who programmed it and gives visibility to the social stereotypes that affect the minds and imagination of minors.

Ethical guidelines for educators on using artificial intelligence (European Commission, 2025) states how necessary it is for educators and teachers to know how to implement ethical educational-didactic design, which using inclusive datasets activate digital critical thinking education from early childhood, with the awareness that AI must be a device at the service of learning and not a tool of discrimination (Annacontini, 2023).

From an educational-didactic perspective, it is necessary to train educators and teachers in agency; learner-teacher interaction; digital literacy and responsibility and trust (Panciroli & Rivoltella, 2023, pp. 84-87), because Artificial Intelligence can be positive for the educational process and can bring about an improvement «as long as the purposes are clear and the data collections transparent» (Redaelli, 2025, p. 69).

The essay aims to demonstrate how through a series of educational strategies (Lorenzoni, 2025) and debiasing techniques (such as debate, evidence-based learning, metacognitive reflection) the impact of biases is reduced, and the mind is trained to critical, reflective, inclusive thinking that is less subject to preconceived judgments.