This pioneering project, supported by the university's "Excellence Initiative – Research University" program, concluded in June.
Terabot, an artificial intelligence (AI)-powered tool, is built to help patients deal with emotions such as anxiety, shame, and anger, and fostering mindfulness.
Unlike traditional pharmacological treatments, Terabot provides a novel approach by engaging patients in interactive dialogues. The system was designed in collaboration with a psychiatrist and psychotherapist, ensuring its relevance in therapeutic contexts.
The project's initial phase involved the use of a computer-generated avatar, controlled by a therapist, to replicate and gradually transform the negative statements heard by patients in their hallucinations.
The method showed significant promise, according to Prof. Artur Janicki from the Faculty of Electronics and Information Technology.
The study, which involved 23 patients, demonstrated notable reductions in the severity of psychotic symptoms and intensity of hallucinations, with effects lasting up to three months.
Building on this success, the team developed the more autonomous Terabot system.
Unlike its predecessor, Terabot operates independently, without direct psychiatrist control. It engages in conversations with patients, recognizing their speech and emotional states, and tailors its responses accordingly. The primary focus of this therapy is to train patients in emotional regulation and mindfulness.
A total of 38 participants were involved in the trial, with many reporting increased openness and communication when interacting with the computer compared to a human therapist.
However, the system is not without its limitations. Challenges remain in refining Terabot's empathetic understanding and response timing. The ongoing research aims to enhance these aspects, making Terabot a more effective and empathetic therapeutic tool.
Terabot is a part of the larger AI and Robotics research initiative at the Warsaw University of Technology. This project, alongside others, received funding under the SzIR-2 research grant competition, showcasing the university's commitment to pioneering AI applications in healthcare.
As the analysis of the study's results continues, the potential of AI in augmenting traditional therapy methods becomes increasingly evident, offering hope for improved patient outcomes.