On this wall I’ll look into different AI-driven technologies that currently exist and how they are or could be utilized in this field. A big part of my research will focus on the implications of bringing AI into mental health care—both the possible benefits and the risks. Psychology is a field that requires a delicate human touch, so introducing AI into this space comes with concerns about ethics, effectiveness, and emotional intelligence on behalf of the AI.

Beyond that, privacy is a huge issue. How do we ensure that AI-powered therapy platforms comply with HIPAA regulations and maintain confidentiality? What happens to all the sensitive patient data that’s collected? Generative AI learns based on all of the input it receives, so where does all your input data go and how do you ensure your personal health data isn't used to train an algorithm? These are the kinds of questions I want to attempt to gain insight on.

I chose this topic because I think AI has the potential to be an incredible resource in this field, especially for those who struggle to access or afford mental health care. But it’s important to examine both its benefits and its potential consequences before completely integrating it into such a sensitive field.

Technology Used: Natural Language Processing , Sentiment Analysis, Machine Learning.

Benefits & Improvements:

• Provides free or low cost 24/7 access to mental health support

• Reduces feeling of personal stigma by offering faceless and anonymous interaction

• Helps bridge the gap for individuals with limited access to therapists (APA, 2025)

  
  

Challenges & Limitations:

• Lacks deep emotional intelligence and true understanding of the issues at hand

• Cannot fully replace human therapists, especially in complex cases

• Raises privacy concerns regarding user data storage and security (SCU Ethics, 2025)

  
  

Personal Insights: This is cool, and AI chatbots certainly have the potential to make mental health care more accessible and immediate. However, as of right now I think that they should be seen as a supplement to people who absolutely need it and cannot otherwise access expert help, rather than a replacement for professional human care.

Technology Used: Deep Learning, Sentiment Analysis, Computer Vision.

Benefits & Improvements:

• Speeds up early diagnosis and intervention

• Reduces human bias' during diagnostic assessments

• Can be integrated into telemedicine platforms for remote screenings (Haque & Rubya, 2023)

Challenges & Limitations:

• Risk of misdiagnosis due to data limitations and lack of human understanding

• Ethical concerns about data privacy and consent (SCU Ethics, 2025)

• AI may reinforce biases if the model was trained on data sets that do not have proper representation (Zhang & Wang, 2024)

  
  

Personal Insights: AI-driven diagnostics have great potential for improving detection and intervention in cases of mental health diagnosis'. Although, there are concerns about accuracy and (again) privacy that must be carefully addressed before this technology gets widespread adoption. It is especially important at this stage to have a human overlook the diagnosis of individuals because some individuals do require prescription medications.

Technology Used: Predictive Analytics, NLP, Machine Learning, Sentiment Analysis.

Benefits & Improvements:

• AI models improve efficiency for suicide prevention hotlines; the most urgent cases receive immediate attention

• AI tools can analyze patterns in text, voice, and online activity to detect potential warning signs of suicide risk (Lejeune et al., 2022)

• AI models can predict suicide risk with up to 80% accuracy based on health records (Zhang & Wang, 2024)

  
  

Challenges & Limitations:

• Ethical concerns about surveillance and privacy

• False positives may lead to unnecessary interventions

• AI models may require further validation to minimize bias and improve reliability (Lejeune et al., 2022)

  
  

Personal Insights: AI’s ability to predict suicide risk based off of online activity and prioritize crisis intervention has the ability to save lives. But, this technology is intrusive and it must be implemented ethically in a way that attempts to avoid privacy violations.

• Rationale: This source provides an overview of AI’s role in mental health, including its benefits, risks, and ethical considerations.

Haque, M. D. R., & Rubya, S. (2023). An Overview of Chatbot-Based Mobile Mental Health Apps: Insights From App Description and User Reviews. JMIR mHealth and uHealth, 11, e44838. Retrieved from https://pmc.ncbi.nlm.nih.gov/articles/PMC10242473/

• Rationale: This examines the effectiveness of AI-driven chatbot applications in mental health care.

Lejeune, J., Brixtel, R., & Bringay, S. (2022). Artificial Intelligence for Suicide Prevention: A Systematic Review of Machine Learning-Based Methods. Frontiers in Psychology, 13, 898827. https://doi.org/10.3389/fpsyg.2022.898827

• Rationale: Examines how AI is used in suicide prevention such as in social media analysis and hotlines.

New York Times. (2025, February 24). AI Therapists: Can Chatbots Replace Your Psychologist? Retrieved from https://www.nytimes.com/2025/02/24/health/ai-therapists-chatbots.html

• Rationale: This article explores the growing trend of AI-driven therapy chatbots and their effectiveness.

Santa Clara University Ethics Center. (2025). AI Therapist & Data Privacy: Ethical Concerns in Mental Health AI. Retrieved from https://www.scu.edu/ethics/focus-areas/bioethics/resources/ai-therapist-data-privacy/ai-therapist-data-privacy.html

• Rationale: This source reviews data privacy concerns and ethics regarding AI mental health tools.

Zhang, Z., & Wang, J. (2024). Can AI Replace Psychotherapists? Exploring the Future of Mental Health Care. Frontiers in Psychiatry, 15, 1444382. Retrieved from https://pmc.ncbi.nlm.nih.gov/articles/PMC11560757/

• Rationale: This article discusses the growing role of AI in the mental health field and debates if it can effectively change the field of traditional therapy.

The ethical considerations while implicating AI in this field have been partially discussed throughout this wall. When considering integrating AI into a field where you are dealing with people's health information it is extremely important to have secure platforms and trustworthy companies running those platforms. Not including backdoors in these platforms as well as implementing federated learning techniques on the training models and being transparent with users are all ways to alleviate ethical concerns.

Some stress that AI could take away jobs, especially for counselors. While it’s true that AI can do tasks like screening for symptoms or guiding people through chatbot conversations, therapy is fundamentally based around human connection. AI is able to coexist with therapists by helping schedule, take notes, or analyze patterns in patient data. In my opinion, despite all the advancement in AI as it currently exists, it will be just a tool, and not a replacement for genuine human connection. In fact, I believe that for some people AI could exacerbate their symptoms. Within the past year a teenager formed a codependent relationship with a chatbot on character.AI and committed suicide over it. In my opinion, we also need to look at case studies such as this when we are incorporating this technology into a field full of vulnerable people. 

At the same time, AI raises concerns about privacy and bias. These systems would collect sensitive mental health data, and without strong protections, there’s a risk of misuse and exploitation. AI also isn’t perfect—it can reflect biases from the data it was trained on, which means it may misinterpret symptoms or be less effective for people from underrepresented backgrounds. 

All in all, AI can definitely be a helpful tool, but right now it can not replace the human connection which is essentially the heart and soul of mental health care. Incorporating AI into this field shouldn’t be about machines taking over therapy—it must be about using it to support the work of people, making mental health care more accessible for everyone.