DHCBT: a digital human psychological counseling system with a CBT multi-agent framework

Academic stress is a major risk factor for mental health problems among college students. Existing online counseling systems often face three limitations: limited user trust in virtual interactions, insufficient adaptation to academic-stress scenarios, and inconsistent professional quality in single-agent counseling. To address these issues, we developed DHCBT, a digital human counseling system based on a hierarchical CBT multi-agent framework.

Coordinated by a meta-agent and specialized expert agents, the framework structures counseling into two stages: establishment of an initial intervention framework and execution of the CBT core process. We evaluated the framework through benchmark testing on CBT-BENCH Level II, response-quality assessment on academic-stress counseling cases, and a controlled user study comparing DHCBT with a text-based chatbot. In benchmark testing, the framework showed strong performance on cognitive distortion and core-belief understanding tasks.

In counseling-response evaluation, it outperformed baseline models in safety, adherence to CBT process, problem comprehension, and empathy. In the user study, DHCBT was associated with lower cognitive load and higher social presence, usability, and intention to continue use than the text-based control. These findings provide preliminary evidence that combining a digital human counseling interface with a CBT multi-agent framework is feasible for supporting college students with mild-to-moderate academic stress.

This work was supported in part by the National Natural Science Foundation of China under Grant 62306044 and in part by the National Social Science and Technology Fund (BGA230254). Beijing University of Posts and Telecommunications, School of Artificial Intelligence, Beijing, 100876, China Litun Tan, Zongze Li, Yitian Huang, Jun Liu & Keliang Chen China Mobile Group Device Co., Ltd., Terminal Company, Beijing, 100053, China The authors declare no competing interests. Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Below is the link to the electronic supplementary material. Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it.

The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.