Toward a Hybrid Model: A Delphi Study of Expert Consensus on Artificial Intelligence in Counseling Supervision

As artificial intelligence (AI) rapidly integrates into counselor education, its role within the relational boundaries of clinical supervision remains ambiguous. This study explores human-AI teaming in supervision through a three-round Delphi design involving an expert panel with international representation (n = 20). Following qualitative and quantitative (median/IQR) iterations, experts reached consensus that AI currently serves best as a support tool for case conceptualization and low-risk simulations. While acknowledging AI’s potential for didactic feedback, panelists expressed critical concerns regarding algorithmic emotional shallowness, inability to grasp cultural nuances, data privacy, and trainee overreliance. Ultimately, findings advocate for a hybrid human-AI supervision framework. This framework explicitly restricts AI to an analytical “assistant” role, reserving the relational core, empathy, and gatekeeping for human professionals. By defining clear boundaries for AI delegation, this study informs future human-computer interaction policies and ethical guidelines in clinical training.