Describe Me Something You Do Not Remember - Challenges and Risks of Exposure Design Using Generative Artificial Intelligence for Therapy of Complex Post-traumatic Disorder Conference: Generative AI and Accessibility Workshop; April 27, 2025; Yokohama, JP

Annalisa Degenhard, Stefan Tschöke, Michael Rietzler, Enrico Rukzio · 2025

Read Describe Me Something You Do Not Remember - Challenges and Risks of Exposure Design Using Generative Artificial Intelligence for Therapy of Complex Post-traumatic Disorder Conference: Generative AI and Accessibility Workshop; April 27, 2025; Yokohama, JP by Annalisa Degenhard, Stefan Tschöke, Michael Rietzler, Enrico Rukzio (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-20796.txt.

Introduction

Annalisa Degenhard and co-authors weigh AI-made images as a therapy tool for complex trauma. Their 2025 paper treats exposure therapy, where patients revisit hard memories in a safe frame to build coping strength. Stock images miss the mark for deeply personal trauma, and custom art costs too much for most clinics.

Generative AI can make personal scenes live in session, with the patient as co-maker. That promise comes with risk: both the illness and the tool breed a sense of lost control. Careful role splits among patient, system, and therapist are urged to keep sessions safe. The paper maps gains, hazards, and open design work for this young field.

Catalog record

CREATORAnnalisa Degenhard, Stefan Tschöke, Michael Rietzler, Enrico Rukzio
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE26 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2505-20796

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Describe Me Something You Do Not Remember - Challenges and Risks of Exposure Design Using Generative Artificial Intelligence for Therapy of Complex Post-traumatic Disorder Conference: Generative AI and Accessibility Workshop; April 27, 2025; Yokohama, JP Annalisa Degenhard, Stefan Tschöke, Michael Rietzler, Enrico Rukzio Abstract . Post-traumatic stress disorder (PTSD) is associated with sudden and uncontrollable, intense flashbacks of traumatic memories. One form of treatment that has been shown to be effective in reducing the severity of trauma-related symptoms is trauma exposure psychotherapy. It involves the controlled recall of traumatic memories to train coping with trauma flashbacks and to allow autobiographical integration of trauma-associated distressing memories. In particular, exposure to visualizations of such memories is an effective approach for successful recall. However, although this form of treatment has been shown to be effective for various types of trauma, it is still only available for a few types of trauma. This is due to a lack of economic solutions for creating individualized exposure visualizations. This is especially true for the treatment of Complex PTSD (CPTSD), as the traumatic memories of this group are mostly individual, and there are no generic visualizations that work for multiple patients. Generative Artificial Intelligence (GAI) offers a flexible, cost-effective solution. GAI not only provides the novel ability to create individualized exposure visualizations during the therapy session, but also allows patients to participate in the visualization process for the first time. With the advent of novel therapeutic perspectives and the potential to improve access to trauma therapy, especially for CPTSD, the use of GAI also presents challenges and exposes the patient to serious risks. The exceptionally high levels of uncertainty and lack of control that characterize both CPTSD and GAI challenge the feasibility of such solutions. To enable safe and effective three-way communication, it is critical to thoroughly understand the roles of the patient, system, and therapist in the context of exposure visualization and how each can contribute to safety. This paper summarizes perspectives, challenges, and risks for the use of GAI in trauma therapy with a focus on CPTSD. 1. Trauma and Dissociation If I were to ask you to describe an incomplete memory in such a way that it could be visualized in a medium unfamiliar to you, and that the aspects most relevant to you could be presented in as much detail as possible, would you be able to do so in an instant? This is one of the main challenges of visualization-based exposure therapy, like Virtual Reality Exposure Therapy (VRET), for Posttraumatic Stress Disorder (PTSD) ( Javidi and Yadollahie 2012a ; Maercker et al. 2018 ) . PTSD is a common health problem often characterized by heightened physiological arousal and constant vigilance, uncontrolled intense flashbacks of traumatic memories, and avoidance of situations that may include potential triggers of such flashbacks ( Shalev et al. 2024 ; Brewin et al. 2025 ; Greenberg et al. 2015 ; Javidi and Yadollahie 2012b ; Javidi and Yadollahie 2012a ) . A common symptom associated with psychological trauma, particularly recurrent interpersonal trauma during childhood, is dissociation ( Stolovy et al. 2015 ) . Dissociation can be roughly described as a coping strategy of the brain that limits the accessibility of cognitive aspects in order to avoid overwhelming experiences ( Chu and Dill 1990 ) . This may include discontinuity of consciousness, memory, and perception ( Dia 2013 ) . Trauma exposure psychotherapy is currently the most promising form of treatment for PTSD following exposure to single and multiple traumatic events ( Hoppen et al. 2024 ) . It focuses on the retrieval of avoided memories in order to train the handling of them ( Brewin et al. 2025 ; Cloitre et al. 2011 ) and to allow an autobiographical integration ( Berntsen and Rubin 2007 ; Smeets et al. 2010 ) . Recall of traumatic memories may be difficult, depending on the patient’s level of dissociation. Approaches include collecting retrospective accounts from traumatized individuals, such as the narrative lifeline approach ( Neuner et al. 2021 ) , post-hoc observations, and provoking and then studying traumatic memories and flashbacks in individuals suffering from PTSD. The retrieval process typically takes time, as traumatic memories are initially accessible mostly through flashbacks or intuitive feelings, and individuals must learn to access them consciously ( van der Kolk and Fisler 1995 ) . However, the effectiveness of exposure therapy depends on successful confrontation with a traumatic stimulus ( Bohus and Vonderlin 2024 ) . This is often challenged by the high levels of avoidance associated with PTSD ( Charlton and Thompson 1996 ) . In particular, approaches that rely on patients imagining the traumatic stimulus may be ineffective because therapists cannot control whether the patient is actually imagining what is being pretended ( Vincelli 1999 ; Difede et al. 2002 ) . One solution is provided by visualization-based approaches such as VRET ( van Meggelen 2019 ; Meggelen et al. 2022 ) . Visualization of traumatic stimuli can at least partially circumvent avoidance. However, this form of treatment depends on the technical and financial feasibility of visualizing the patient’s traumatic stimuli. VRET has shown great feasibility for combat-related trauma, as traumatic memories have been shown to be relatively similar across individuals, and visualizations have been shown to be effective for multiple individuals ( Rizzo et al. 2021 ) . Complex posttraumatic stress disorder (CPTSD) can result from exposure to multiple traumatic events ( Wilson 2004 ; Cloitre et al. 2014 ) . Unlike monotraumatized individuals, the traumatic stimuli described by affected individuals are often highly individualized, and generalization across patients is difficult ( Wilson 2004 ) . CPTSD is also associated with greater levels of dissociation, which makes the recall process more difficult and often follows the principle of trial and error ( Brewin et al. 2025 ; Hyland et al. 2020 ) . A common cause of CPTSD is childhood trauma. 2. Accessibility of Treatment for CPTSD Visualization-based trauma exposure has already been shown to be an effective therapy for childhood trauma in both mono- and complexly traumatized patients using individualized exposure therapy ( Chu and Dill 1990 ; Hoppen et al. 2024 ) . Psychopharmacological treatment with medication is not currently available ( Leichsenring et al. 2024 ) . Given the prevalence of CPTSD ( Maercker et al. 2022 ; Karatzias et al. 2017 ) and the current lack of other effective forms of treatment, psychotherapy with trauma exposure still seems to be the most promising perspective. However, solutions are needed that allow for a customizable and economical visualization of traumatic memories to improve accessibility ( M. Shae Nester and Brand 2022 ) . Enabling flexible, economic visualization, generative artificial intelligence (GAI) seems to provide a solution. Through exceptional visualization speed, for the first time, the creation of individualized exposure visualizations within the therapy session is enabled. This could solve the problem of describing an unclear memory when not knowing what is relevant for the system to successfully visualize a memory. As an iterative design process becomes feasible, patients may actively participate in the visualization process for the first time, which may improve the efficacy of a memory-oriented therapy approach. 3. CPTSD-Related Risks & Challenges in Exposure Therapy The feasibility of

Read the complete text