VTplus engages with artificial intelligence (AI) where its use is clinically meaningful and controllable within a medical technology context: in the analysis of multimodal data, the structuring of therapeutic feedback, the adaptive configuration of virtual scenarios and the support of therapeutic decision-making processes.
In addition, VTplus develops and uses AI-assisted tools for internal knowledge and quality management. These include, in particular, locally operated agent-based systems for source-grounded querying, review and curation of company knowledge, requirements defined in the quality management manual, workflows and regulatory knowledge bases.
In this context, AI is understood as a supporting technology embedded in professional, clinically and regulatorily controlled processes. It does not replace guideline-concordant psychotherapeutic or medical treatment, but may help structure information, make complex data more accessible and support controlled development and quality processes.
Artificial Intelligence in VR Therapy
Artificial intelligence (AI) can complement virtual reality in therapy and research wherever complex data are collected during VR-assisted exercises, structurally analysed and made usable for therapeutic decision-making. The focus is not on replacing therapeutic treatment, but on supporting therapists through data-driven assistance functions.
VR-assisted behavioural and exposure exercises generate different sources of data. These include movements within the virtual environment, gaze direction, interactions with virtual persons or objects, speech, self-reports and — depending on the system architecture — physiological parameters. AI methods can help integrate such multimodal data, identify patterns and present therapeutically relevant information in a clear and structured form.
One possible field of application is the adaptive control of virtual situations. Virtual environments can be designed in such a way that difficulty, stimulus intensity or social feedback can be modified in a controlled manner. This is particularly relevant for exercises addressing social anxiety, public-speaking situations, job interviews or other interactive stress-related situations in which virtual persons, audiences or social responses can be used in a targeted way.
In the BMBF-funded research project OPTAPEB, a VR demonstrator for agent-guided, patient-centred emotion regulation was developed. The system combined virtual social interaction and public-speaking scenarios with multimodal data acquisition, data fusion, virtual agents and AI-assisted intervention suggestions. The application was implemented on the basis of a VT+ExpoCart3 and evaluated in clinical as well as applied research contexts. It was subsequently also tested in vocational rehabilitation as part of the KI.ASSIST project.
The VTplus VR Platform enables the modular use of AI-based analysis and control processes. The platform connects VR simulation software, interactive virtual environments, therapeutic control functions, interfaces to measurement systems and functions for documentation and evaluation. AI-assisted modules can complement this architecture, for example through data-based feedback, adaptive scenario control or support in the evaluation of therapeutic and research-related workflows.
AI-assisted VR systems remain embedded in a professional, clinical and regulatory overall concept. They do not replace guideline-concordant psychotherapeutic or medical treatment. Their appropriate use lies in supporting professional users, the structured implementation of VR exercises and the scientific further development of digital therapy and research systems.
