Personality, identity, and Artificial Intelligence: a grand challenge

activities are offloaded to AI and people are hired, fired, and monitored by AI (e.g., Bankins et al., 2023).Similarly, social relationships are increasingly structured around online communication, smartphone apps, and AI technology. Research on human-machine systems shows that even peripheral AI agents can shape collective outcomes. Importantly, these outcomes arise from interdependent human-human and human-machine interactions, not isolated actors (Tsvetkova et al., 2024). This means that a new challenge for social psychology is to treat AI as a social agent and theorize agency, trust, and moral influence in multi-agent (hybrid) systems. Because these hybrid systems are embedded within broader cultural ecologies, identity processes are simultaneously shaped by evolving social norms and collective narratives. Online identities are shaped by dynamic cultural trends, and how such trends play out will likely differ according to pre-existing cultural differences, within and between nations.Research areas are also differentiated by their focus on the individual versus the social construction of identity. At the individual level, research builds on existing personality trait models, human-computer interaction (HCI) research, and applied cognitive psychology to investigate how people vary in their usage of AI-powered technologies and in their reactions to them. Beyond individual differences in human-AI interaction, personality and social psychology must address identity. At the micro level, AI is now embedded in everyday settings and shapes how identities are activated and regulated through feedback. At the macro level, AI forms part of the sociotechnical infrastructure through which collective meaning, moral values, and group boundaries are constructed, negotiated and refined.Research on personality, identity, and AI is thriving but it is often scattered and disconnected. Here, we aim to promote the coherence of research by identifying major research areas differentiated by the twin axes of (1) granularity vs. high-level, 'big picture' issues and (2) individual differences versus social identity. Within this 2 × 2 scheme, we briefly define key research questions and new approaches to understanding personality in the digital age.Established personality dimensions such as internet anxiety, technophobia/philia, and computer selfefficacy shape attitudes, emotions, and behaviors towards conventional computer systems (Matthews et al., 2021). However, fundamental differences between conventional digital systems and modern AIs will change the role of personality in digital interactions. These include the vast cognitive capabilities of AI, social agency and natural language communication, the assumption of a consistent humanlike persona, and a relationship history with the user. New scales for constructs such as trust in AI, humanlikeness, and social presence are proliferating (Esterwood et al., 2021), but scale development is at a Wild West stage. Quality and evidence for validity of scales vary greatly and there is no overarching psychometric model to support integration of research findings.Research on the sources, development, and malleability of individual differences is also lacking.Concerns about effects of smartphone use and social media on child development (Haidt, 2024) often neglect nuances including individual differences in vulnerability to harmful impacts (Matthews, Herzog & Esau, in press). Longitudinal studies of development of attitudes to AI are needed, in both children and adults; current research on new personality constructs is typically cross-sectional. Understanding the role of established, biologically-based traits such as the Big Five is essential, but some novel constructs, such as online personas, are less stable and more malleable than conventional traits (Olivero et al., 2020).A third research focus is the consequences of attitudes and emotions towards AI. What personality factors influence offloading significant life-decisions to AI, treating an AI as a friend and confidant, and allowing an AI agent to manage one's social media interactions? Such questions can be addressed in both experimental studies and within prospective studies of how individual differences interact dynamically with AI usage.The sociotechnical perspective on impacts of AI positions the individual within interacting technological, organizational, and cultural systems (Matthews et al., 2025;Yu, Xu & Ashton, 2023), suggesting multiple goals for personality research. First, AI introduces novel threats such as uncertainty over the basis for AI judgements, social stressors associated with humanlike interactions with AIs, and loss of self-esteem as AIs exceed humans in cognitive capabilities and decision authority (Matthews et al., 2024). There is an overarching threat to the person's connection to reality or epistemic rationality. Maintaining wellbeing and a sense of cohesion depends not only on basic and AI-linked traits, but also on socioeconomic status and the psychosocial environment (Brunner & Marmot, 2006). How will changes in social organization driven by AI interact with traits to produce stress at individual and group levels? Second, AI enables both benefits and harms (e.g., Bond, Ennis & Mulvenna, 2025). AI can potentially improve quality of life by freeing people from mental drudgery, providing decision support based on a vast knowledge base, and powering therapeutic interventions. On the downside, its knowledge base is contaminated by human biases, including race and gender bias, its outputs are frequently unexplained and unverified, it is indifferent to privacy, and it is readily utilized for malicious purposes such as cybercrime and disseminating misinformation. In organizations, AI enables practices such as intrusive surveillance and algorithmic hiring and firing that dehumanize work relationships. Personalized alignment of AI to match the user's personality, skills, and values can enhance its usefulness and widen access, but also risks infringing privacy and reinforcing bias (Kirk, Vidgen, Röttger & Hale, 2024).Personalized alignment is more than a technical design issue. It raises difficult questions about how personalization should be bounded, and who decides on bounding principles (Kirk et al., 2024).Third, usage and impacts of AI reflect fluid and sometimes contested cultural norms, such as what level of government regulation of AI should be implemented. There is also a growing literature on crosscultural differences in attitudes to AI and how they vary with cultural dimensions such as individualismcollectivism (Barnes, Zhang & Valenzuela, 2024). Another research challenge is that AI domination of life is likely to lead to new human cultures without historic parallel, requiring emic perspectives on attributes of personality bound to digitalized culture to complement universal, etic models (Matthews et al., 2021).AI systems do not merely deliver information but increasingly function as interactive social actors, as they respond, adapt, and mirror users, creating feedback loops that contribute to the activation, validation, and regulation of social identities (Metzler & García, 2023;Pedreschi et al., 2025). Through these interactive processes, AI-mediated environments curate, amplify, and stabilize collective narratives that provide meaning, relevance, and moral orientation to group memberships (Bliuc et al. 2024;Sartori & Theodorou, 2022). These developments pose a theoretical challenge for our field because classic theories in social psychology assume that identities are shaped through interaction with other people, groups, and social institutions (Spears, 2021). However, AI introduces a novel "interaction partner" one that is highly responsive, personalized, scalable, and opaque, yet embedded in everyday communication, community participation, and information access. Existing theoretical models are not well-equipped to explain how identity dynamics operate when group-relevant narratives are curated or generated by adaptive systems. A central research question is whether AI-driven adaptation stabilizes social identities (perhaps by providing coherence, validation, and belonging) or narrows them, by reinforcing exclusionary narratives and reducing identity flexibility. A related question concerns wellbeing -for example, under what conditions do AI-mediated collective narratives support meaning, relevance, and psychological resilience, and when do they contribute to dependency or entrenchment? AI-augmented online communities provide an important test case. In contexts such as addiction recovery or mental health support groups, AI systems may scaffold positive identity change by supporting new recovery narratives (Li et al., 2023;Thakkar et al., 2024), but they may also reinforce static or dependent identities, with other risks associated as well (Elyoseph et al., 2024). Importantly, issues such as trust, vulnerability to misinformation, and polarization should be analyzed not only as cognitive biases but as identity-regulatory processes operating through perceived group alignment and collective meaning-making (Efstratiou & De Cristofaro, 2022).AI systems in conjunction with online media are increasingly becoming part of the infrastructure through which collective identities are constructed, represented, and evaluated. They can mediate how individuals and groups are seen by others, how collective narratives circulate and evolve, and how meaning and moral values are articulated and contested (Gerbaudo, 2022). Psychology currently lacks models of identity and agency for conditions in which aspects of identity may be delegated to nonhuman agents (e.g., use of avatars, digital twins, etc.), personal and interaction data become persistent identity traces, and social recognition is mediated by algorithmic systems rather than interpersonal or institutional processes (Bonnefon et al., 2023). These developments raise important questions about how collective meaning and moral order are sustained, both now and in the future. One risk concerns identity delegation and representation. When AI systems generate profiles, narratives or predictions about individuals or groups, they relocate control over identity representation to system designers and other (ambiguous) operators, raising questions about agency, power, and voice. A second risk concerns collective meaning and moral values. Attitudes toward AI, privacy, and technological change often reflect group-based values and shared (ideologically bound) narratives rather than being driven by deliberative ethical reasoning, with implications for intergroup relations and polarization. A third risk concerns power and inequality, as AI-mediated systems may differentially amplify, normalize, constrain, or erase particular identities. If personality and social psychology fail to theorise identity, meaning, and moral boundaries under AI mediation, we will lack empirically grounded and theoretically informed frameworks to detect, evaluate, and mitigate the potential engineering of identity by platforms, corporations, and state actors.The AI revolution has stimulated a creative ferment of new research in personality and social psychology. We advocate a more systematic approach than currently exists, cohering around welldefined research questions related to individual differences in interactions with AI and to negotiation of identity in communities mediated by AI. Table 1 summarizes some major topics for research we have briefly discussed. These are not intended to be exhaustive but illustrate some directions open to focused investigation. There is also a need for multidisciplinary approaches including perspectives from computer science, neuroscience, communication science, and sociology. We encourage researchers interested in these issues to consider the Personality and Social Psychology section of Frontiers in Psychology as an outlet for their work.