Desafíos actuales de la Inteligencia Artificial

64 Desafíos actuales de la Inteligencia Artificial Figure 2: Share of CNIL fines to entities using AI systems, compared to total fines imposed (2018-2023) Source: CNIL Interestingly, a closer look at the nature of the data processing activities at hand shows that fines particularly affected entities heavily relying on personal data for their AI-powered services. In fact, the overwhelming majority of fines where imposed on companies for their AI-related activities (see Figure 2 above). The CNIL has fined companies using AI systems in a number of domains, namely: data security, personalised advertising, facial recognition, and data brokerage. Table 1 below offers an overview of the main decisions issued by the CNIL related to activities involving the use of AI since the GDPR became applicable. Although the overview is non-exhaustive, it con- tains the most important decisions based on the fines involved. It is also representative of the focus given by the CNIL in its enforcement action, for instance with regards to personalised advertising.