Wednesday 22 July 14:05 - 15:35

Hall: 16 - Room 13 SA

Chair and Presenter: Hogrefe Antonia

Division: Division 2: Psychological Assessment and Evaluation

Artificial intelligence (AI) has rapidly entered the field of psychological assessment, reshaping long-standing practices in both applied and clinical contexts. While offering unprecedented opportunities, AI also poses significant challenges for the validity and integrity of established assessment methods. This symposium brings together five diverse contributions that illustrate the dual role of AI: as both a threat to traditional assessments and a driver of innovative solutions. The first presentation addresses one example of the most pressing risks—AI-assisted faking in personality assessment. Generative AI can produce highly convincing response patterns, raising serious concerns for fairness and validity in personnel selection, but also pointing to new design strategies for prevention. A second contribution highlights how AI can reduce human subjectivity in neuropsychological testing by introducing fully automated scoring of the Rey Complex Figure Test, with promising implications for early detection of cognitive impairment. The third presentation explores AI as an effective and valid alternative to human expert ratings in situational assessments, comparing self-report, expert, and AI-based ratings of open responses. The fourth contribution critically examines whether AI can simulate normative data, using ChatGPT to generate norms for a verbal reasoning task—providing early evidence of both potential and current limitations. Finally, the MAPA project demonstrates how AI-powered natural language processing can be used to assess psychopathology, offering validated digital tools for OCD and anxiety that integrate seamlessly into clinical practice. Together, these perspectives show that AI represents a "next-level challenge" for psychological assessment—undermining validity in some contexts, while offering novel and paradigm-shifting pathways to overcome existing limitations. By addressing both risks and opportunities, this symposium invites applied psychologists to critically reflect on how AI can be integrated responsibly, ensuring assessments remain scientifically rigorous, fair, and clinically useful.