Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
51 tokens/sec
GPT-4o
60 tokens/sec
Gemini 2.5 Pro Pro
44 tokens/sec
o3 Pro
8 tokens/sec
GPT-4.1 Pro
50 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Challenges of Large Language Models for Mental Health Counseling (2311.13857v1)

Published 23 Nov 2023 in cs.CL, cs.AI, and cs.HC

Abstract: The global mental health crisis is looming with a rapid increase in mental disorders, limited resources, and the social stigma of seeking treatment. As the field of AI has witnessed significant advancements in recent years, LLMs capable of understanding and generating human-like text may be used in supporting or providing psychological counseling. However, the application of LLMs in the mental health domain raises concerns regarding the accuracy, effectiveness, and reliability of the information provided. This paper investigates the major challenges associated with the development of LLMs for psychological counseling, including model hallucination, interpretability, bias, privacy, and clinical effectiveness. We explore potential solutions to these challenges that are practical and applicable to the current paradigm of AI. From our experience in developing and deploying LLMs for mental health, AI holds a great promise for improving mental health care, if we can carefully navigate and overcome pitfalls of LLMs.

PDF HTML Abstract
Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize File Document Download Save Streamline Icon: https://streamlinehq.com PDF File Document Download Save Streamline Icon: https://streamlinehq.com Markdown Bookmark Chat Bubble Oval Streamline Icon: https://streamlinehq.com Chat (Pro)
Definition Search Book Streamline Icon: https://streamlinehq.com
References (62)
  1. doi:10.1016/s2215-0366(20)30347-3.
  2. doi:10.1016/s0140-6736(21)02143-7.
  3. doi:10.1016/s2215-0366(21)00395-3.
  4. doi:10.2471/blt.20.273383.
  5. arXiv:2103.00020, doi:10.48550/ARXIV.2103.00020.
  6. OpenAI, Gpt-4 technical report (Mar. 2023). arXiv:2303.08774, doi:10.48550/ARXIV.2303.08774.
  7. arXiv:1810.04805, doi:10.48550/ARXIV.1810.04805.
  8. doi:10.31234/osf.io/cuzvr.
  9. arXiv:2203.15144, doi:10.48550/ARXIV.2203.15144.
  10. doi:10.1109/bigcomp57234.2023.00097.
  11. arXiv:2309.15461, doi:10.48550/ARXIV.2309.15461.
  12. arXiv:2307.11991, doi:10.48550/ARXIV.2307.11991.
  13. doi:10.2196/15708.
  14. doi:10.18653/v1/n19-1423.
  15. doi:10.18653/v1/2020.acl-main.703. URL https://aclanthology.org/2020.acl-main.703
  16. doi:10.1162/neco.1997.9.8.1735.
  17. Common Crawl, https://commoncrawl.org/ (Nov. 2023).
  18. arXiv:2101.00027, doi:10.48550/ARXIV.2101.00027.
  19. doi:10.1126/science.aal4230.
  20. doi:10.1073/pnas.2313790120.
  21. doi:10.2196/40242.
  22. doi:10.18653/v1/2020.acl-main.463.
  23. arXiv:2208.02957, doi:10.48550/ARXIV.2208.02957.
  24. arXiv:2308.15399, doi:10.48550/ARXIV.2308.15399.
  25. doi:10.1145/3571730.
  26. doi:10.1101/2023.03.16.23287316.
  27. doi:10.3350/cmh.2023.0089.
  28. doi:10.1109/ichi.2018.00039.
  29. doi:10.18653/v1/2020.acl-main.442.
  30. doi:10.3389/frai.2023.1229805.
  31. doi:10.1145/3287560.3287596.
  32. arXiv:1706.02744, doi:10.48550/ARXIV.1706.02744.
  33. doi:10.2196/38799.
  34. doi:10.1136/amiajnl-2011-000089.
  35. doi:10.1038/s42256-019-0048-x.
  36. doi:10.18653/v1/P19-3007. URL https://www.aclweb.org/anthology/P19-3007
  37. doi:10.1055/s-0040-1701980.
  38. doi:10.1176/appi.ajp.2014.14030423.
  39. doi:10.2196/28003.
  40. doi:10.1016/j.beth.2021.09.007.
  41. doi:10.1093/jamiaopen/ooac039.
  42. arXiv:2106.09685, doi:10.48550/ARXIV.2106.09685.
  43. arXiv:2311.03285, doi:10.48550/ARXIV.2311.03285.
  44. arXiv:2310.18547, doi:10.48550/ARXIV.2310.18547.
  45. doi:10.18653/v1/d17-1323.
  46. doi:10.18653/v1/2020.findings-emnlp.301.
  47. doi:10.1007/s43681-023-00289-2.
  48. arXiv:2204.02311, doi:10.48550/ARXIV.2204.02311.
  49. arXiv:2305.10403, doi:10.48550/ARXIV.2305.10403.
  50. doi:10.1038/s41586-023-06291-2.
  51. doi:10.1145/3461702.3462540.
  52. doi:10.2105/ajph.91.11.1869.
  53. doi:10.1037/0033-2909.129.5.674.
  54. doi:10.2217/fmai-2023-0011.
  55. doi:10.1056/nejmsr2214184.
  56. doi:10.1038/s43856-023-00370-1.
  57. doi:10.1001/jamainternmed.2023.1838.
  58. doi:10.1038/s41746-023-00852-5.
  59. doi:10.3389/fmed.2022.990604.
  60. doi:10.1126/science.217.4566.1237.
  61. doi:10.1002/aaai.12036.
  62. doi:10.15252/embr.201949177.
User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (3)
  1. Neo Christopher Chung (13 papers)
  2. George Dyer (2 papers)
  3. Lennart Brocki (11 papers)
Citations (9)
View on Semantic Scholar