Enabling access or automating empathy? Using chatbots to support GBV survivors in conflicts and humanitarian emergencies
Bibliographic record
Abstract
Abstract Interest in the use of chatbots powered by large language models (LLMs) to support women and girls in conflicts and humanitarian crises, including survivors of gender-based violence (GBV), appears to be increasing. Chatbots could offer a last-resort solution for GBV survivors who are unable or unwilling to access relevant information and support in a safe and timely manner. With the right investment and guard-rails, chatbots might also help treat some symptoms related to mental health and psychosocial conditions, extending mental health and psychosocial support (MHPSS) to crisis-affected communities. However, the use of chatbots can also increase risks for individual users – for example, generating unintended harms when a chatbot hallucinates or produces errors. In this paper, we critically examine the opportunities and limitations of using LLM-powered chatbots 1 that provide direct care and support to women and girls in conflicts and humanitarian crises, with a specific focus on GBV survivors. We find some evidence in the global North to suggest that the use of chatbots may reduce self-reported feelings of loneliness for some individuals, but we find less evidence on the role and effectiveness of chatbots in crisis counselling and treating depression, post-traumatic or somatic symptomology, particularly as it relates to GBV in emergencies or other traumatic events that occur in armed conflicts and humanitarian crises. Drawing on key expert interviews as well as evidence and research from adjacent scholarship – such as feminist AI, trauma treatment, GBV, and MHPSS in conflicts and emergencies – we conclude that the potential benefits of GBV-related, AI-enabled talk therapy chatbots do not yet outweigh their risks, particularly when deployed in high-stakes scenarios and contexts such as armed conflicts and humanitarian crises.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.006 | 0.029 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.002 | 0.002 |
| Scholarly communication | 0.003 | 0.003 |
| Open science | 0.001 | 0.004 |
| Research integrity | 0.002 | 0.001 |
| Insufficient payload (model declined to judge) | 0.011 | 0.002 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".