The possibility that artificial intelligence systems may develop a form of personal identity depends on whether social cognition can be attributed to them—that is, the capacity to understand and participate in social interactions. This complex philosophical question was the focus of a lecture delivered by Petra Vidaković Cvetković, which examined the relationship between social cognition and artificial agents.
The lecture, titled “Social Cognition and the Possibility of Personal Identity in AI Agents,” focused on analyzing the relationship between social cognition and artificial agents, with particular attention to systems based on deep neural networks. The central thesis was that the modeling and implementation of social-cognitive capacities have far-reaching implications for the status of personhood in artificial agents, and that these capacities may provide a relevant framework for exploring a minimal conception of personal identity in the context of artificial intelligence.
In the first part of the lecture, the conceptual framework of social cognition developed by Cameron Buckner in From Deep Learning to Rational Machines was presented and analyzed. Special attention was given to the relationship between deep learning, rational reasoning, and social interaction, as well as to the relevance of these elements for understanding the behavior of artificial agents.
The lecture then introduced the complementary perspective of Anna Strasser, whose minimalist approach to social cognition allows for theoretical integration with Buckner’s position. On the basis of this integration, the possibility of attributing personhood to artificial agents capable of minimal forms of social cognition was examined.
The theoretical foundation of this integrated approach is the criterion of psychological continuity in diachronic personal identity. In this context, particular emphasis was placed on the problem of “para-personhood” as a specific case of the broader problem of personal identity persistence, as formulated by James DiGiovanna. This problem highlights the challenges faced by proponents of psychological continuity in accounting for identity over time, especially in the case of artificial agents.
In the concluding part of the lecture, an anticipated objection to this position was addressed, along with a reasoned response aimed at defending the possibility of a minimal conception of personal identity in artificial agents. The lecture thus opened further space for discussion on the boundaries of personhood, identity, and social agency in the era of advanced artificial intelligence systems.
The lecture was delivered by Petra Vidaković Cvetković, a research trainee at the Center for Philosophy, as part of a lecture series featuring newly appointed colleagues at the Institute of Social Sciences.
