Skip to content
AI News HubLIVE
Original source2 min read

CrisisBench: an evaluation suite for crisis counseling AI

Summary

CrisisBench is a privacy-preserving evaluation suite for crisis counseling AI, featuring single- and multi-turn tasks and simulated patient LLMs. Evaluations show all systems underperform human counselors on excessive empathy and questioning, though they excel at core structural competencies.

SourcearXiv Computational LinguisticsAuthor: Akshay Swaminathan, Filip Ramazan, Sharang Phadke, Kevina Wang, Ivan Lopez, Shaked Peleg Azzam, Gloria Ye, Chastin Chung, William Wang, Stephanie Stoll, Ivy Pham, Rebecca Hurwitz, Shreya Shah, Samuel Chuang, Divyanjali Verma, Abhay John, Ehsan Adeli, Nigam H. Shah
CrisisBench: an evaluation suite for crisis counseling AI
Report an error

The correction channel is not available yet. You can copy the article reference below for later.

Correction instructions
Read article

[Submitted on 11 Sep 2026]

Title:CrisisBench: an evaluation suite for crisis counseling AI

View a PDF of the paper titled CrisisBench: an evaluation suite for crisis counseling AI, by Akshay Swaminathan and 17 other authors

View PDF

Abstract:Background: Crisis helplines are essential to mental health systems but are often understaffed and underfunded. LLMs could support counselors, but existing benchmarks do not capture the escalation, risk assessment, and de-escalation central to crisis conversations, while privacy restrictions limit use of real transcripts. Methods: We present CrisisBench, a privacy-preserving suite for single- and multi-turn crisis counseling evaluation. It includes 1,085 single-turn tasks from 200 de-identified conversations and 1,185 patient LLMs derived from hotline metadata. Patient models and a 23-item LLM judge were calibrated against expert counselors. Results: We evaluated five GPT models plus draft-then-edit and retrieval-augmented generation. All systems underperformed human counselors on excessive empathy and questioning in single-turn evaluation. The rubric achieved human-human Krippendorff's $\alpha$ = 0.51 and LLM-human $\alpha$ = 0.54 (Pearson r = 0.80). In multi-turn evaluation, retrieval-augmented generation scored highest overall (0.961), with all systems within 0.058. Microskills and style/tone were near ceiling, while productivity and safety/risk accounted for most between-system variation. Counselors rated patient LLMs realistic (86%) and useful (92%). Conclusion: CrisisBench enables scalable, privacy-preserving evaluation of crisis counseling. Current systems perform strongly on core structural competencies but still diverge from experts on important response-level failure modes. We release the rubrics, judge prompts, and evaluation tasks.

Subjects:

Computation and Language (cs.CL)

Cite as: arXiv:2611.21622 [cs.CL]

(or arXiv:2611.21622v1 [cs.CL] for this version)

https://doi.org/10.48550/arXiv.2611.21622

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Filip Ramazan [view email] [v1] Fri, 11 Sep 2026 06:42:30 UTC (1,585 KB)

Full-text links:

Access Paper:

View a PDF of the paper titled CrisisBench: an evaluation suite for crisis counseling AI, by Akshay Swaminathan and 17 other authors

View PDF

view license

Additional Features

Audio Summary

Current browse context:

cs.CL

new | recent | 2026-11

Change to browse by:

cs

References & Citations

SciX

Google Scholar

Semantic Scholar

Loading...

Data provided by:

Bibliographic Tools

Bibliographic and Citation Tools

Bibliographic Explorer Toggle

Bibliographic Explorer (What is the Explorer?)

Connected Papers Toggle

Connected Papers (What is Connected Papers?)

Litmaps Toggle

Litmaps (What is Litmaps?)

scite.ai Toggle

scite Smart Citations (What are Smart Citations?)

Code, Data, Media

Code, Data and Media Associated with this Article

alphaXiv Toggle

alphaXiv (What is alphaXiv?)

Links to Code Toggle

CatalyzeX Code Finder for Papers (What is CatalyzeX?)

DagsHub Toggle

DagsHub (What is DagsHub?)

GotitPub Toggle

Gotit.pub (What is GotitPub?)

Huggingface Toggle

Hugging Face (What is Huggingface?)

ScienceCast Toggle

ScienceCast (What is ScienceCast?)

Demos

Demos

Replicate Toggle

Replicate (What is Replicate?)

Spaces Toggle

Hugging Face Spaces (What is Spaces?)

Spaces Toggle

TXYZ.AI (What is TXYZ.AI?)

Related Papers

Recommenders and Search Tools

Link to Influence Flower

Influence Flower (What are Influence Flowers?)

Core recommender toggle

CORE Recommender (What is CORE?)

Author

Venue

Institution

Topic

About arXivLabs

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)

Key points and analysis

Article intelligence

InvestorsAdvanced

Key points

  • Includes 1,085 single-turn tasks and 1,185 patient LLMs from 200 de-identified conversations
  • All systems underperformed human counselors on excessive empathy and questioning in single-turn evaluation
  • Retrieval-augmented generation scored highest overall (0.961) in multi-turn evaluation, with all systems within 0.058
  • Rubric achieved human-human α=0.51 and LLM-human α=0.54 (Pearson r=0.80)

Highlights and analysis are generated automatically and may contain errors. Check the original source.