AI Listens but Does Not Understand
Shreya G.
July 03, 2025
As an undergrad student, stress becomes a baseline. With back-to-back exams, a job, a social life, exercise, and adequate sleep in between, it is easy to become overworked and overwhelmed. Even finding time for care seems like another task on the list. That’s when the idea of optimization kicks in.
And what is more efficient— what better way is there to cut as many corners as possible than to use AI?
Feeling this exact sentiment, I asked ChatGPT:
“I’m feeling extremely burnt out right now. How do I keep up with everything?”
The response was well-structured. It encouraged self-reflection and recommended breathing exercises and journaling prompts. On its surface, it was helpful. But something was missing.
It did not ask thoughtful follow-up questions. There was no deeper reflection. No recognition of cultural values or identity. And certainly no understanding of what it means to navigate mental health as a child of immigrants.
AI is changing how we approach mental health care. A quick search for “AI therapy” and you’ll find apps like Woebot, Wysa, and yes, even ChatGPT, used by millions to manage stress and anxiety.
But what I got back didn’t feel personal. It felt generic.
AI therapy can give advice, but it can’t truly understand.
Where AI Therapy Falls Short
It is true that AI-based mental health tools can lead to short-term improvements in depression and distress (1). However the keyword here is short-term. When it comes to lasting change, deeper context, cultural nuance, and human connection are needed.
Here are my concerns about AI therapy:
- Little impact on long-term well-being: While some people may experience temporary relief, AI therapy has not been shown to significantly improve long-term psychological well-being (1).
- No cultural understanding: AI cannot grasp the complexities of South Asian family dynamics and cultural expectations (2).
- Surface-level help: AI systems can never replicate the qualities of human therapists: genuine empathy, ethical judgment, and the ability to interpret body language and nonverbal cues (3).
- Dangerous reinforcement: A recent study by Stanford found that large language models like ChatGPT stigmatize some mental health conditions over others and provide unsafe or inappropriate responses in certain scenarios surrounding trauma (4).
- No processes for accountability: The majority of chatbots are unregulated, which can put vulnerable individuals at risk. The FDA has not approved AI as a method for diagnosing and treating mental health conditions (5).
- Inherent bias: AI is based on predominantly Western datasets, so it will inherently overlook and misunderstand non-Western cultures and dynamics (6).
What Sukoon Offers Instead
At Sukoon, we believe care that is deeply personal, grounded in culture, and built on connection is more necessary than ever. Our approach recognizes the weight of identity and lived experiences in ways that algorithms simply cannot.
References
Li, H., Zhang, R., Lee, Y. C., Kraut, R. E., & Mohr, D. C. (2023). Systematic review and meta-analysis of AI-based conversational agents for promoting mental health and well-being. npj Digital Medicine, 6(1), Article 236. https://doi.org/10.1038/s41746-023-00979-5
Naeem, F., Khan, N., Sohani, N., Safa, F., Masud, M., Ahmed, S., Thandi, G., Mutta, B., Kasaam, A., Tello, K., Husain, M. I., Husain, M. O., Kidd, S. A., & McKenzie, K. (2024). Culturally adapted cognitive behaviour therapy (CaCBT) to improve community mental health services for Canadians of South Asian origin: A qualitative study. Canadian Journal of Psychiatry, 69(1), 54–68. https://doi.org/10.1177/07067437231178958
Zhang, Z., & Wang, J. (2024). Can AI replace psychotherapists? Exploring the future of mental health care. Frontiers in Psychiatry, 15, 1444382. https://doi.org/10.3389/fpsyt.2024.1444382
Wells, S. (2025, June 11). Exploring the dangers of AI in mental health care. Stanford Institute for Human-Centered AI. Retrieved June 2025, from https://hai.stanford.edu/news/exploring-the-dangers-of-ai-in-mental-health-care
American Psychological Association. (2025, March). Using generic AI chatbots for mental health support. APA Services. Retrieved June 2025, from https://www.apaservices.org/practice/business/technology/artificial-intelligence-chatbots-therapists
Norori, N., Hu, Q., Faraci, F. D., & Tzovara, A. (2021). Addressing bias in big data and AI for health care: A call for open science. PLOS Digital Health, 1(10), e0000123. https://doi.org/10.1371/journal.pdig.0000123