AI and mental health: Curiosity, caution and the human connection
By Miriam Mintz, Clinical Supervision Manager, Able Futures: Access to Work Mental Health Support Service.
Artificial intelligence (AI) is changing how we live and work and the mental health sector is no exception. Whether we feel excited, sceptical or overwhelmed, AI is already a part of our professional and personal lives.
It offers significant opportunities but also raises important questions about ethics, trust, professional judgement and human connection. For me, the question is no longer whether AI has a place in mental health: it is already here. What matters is how we engage with it thoughtfully, with curiosity rather than fear.
The British Association for Counselling and Psychotherapy (BACP) has published its updated Ethical Framework for the Counselling Professions 2026, which comes into effect in November. It now explicitly addresses AI, reinforcing the importance of confidentiality, informed consent, transparency and professional accountability.
This is welcome. Technology is advancing faster than regulation and while existing UK legislation (such as GDPR) provides some safeguards, there is a lack of AI-specific legislation or regulation.
Professional guidance is therefore essential, but so is our willingness to think critically and take responsibility for how technology is used in practice.
Opportunity and uncertainty
AI has genuine potential to improve mental health support. It may reduce administrative burdens, improve accessibility, support learning and overcome language barriers. For those struggling to access traditional services, it may offer another route to information and support.
However, we must distinguish between using AI to support individuals or professionals and allowing it to replace the human elements of care. At the heart of counselling, psychotherapy and mental health coaching is the therapeutic relationship. Trust, empathy, authenticity and psychological safety are fundamental.
AI can produce responses that sound empathic, but this is not the same as human empathy or the experience of being genuinely understood. Mental health work involves noticing what remains unsaid, sitting with uncertainty and gently challenging assumptions when appropriate. These are deeply human skills that technology cannot easily replicate.
AI is already showing up in unexpected ways
As AI becomes more embedded in everyday life, we’re beginning to encounter it in situations that raise new and sometimes unexpected questions about trust, authenticity and human connection.
I have heard of an AI presence joining online sessions, appearing visually as a user. You can imagine that the individuals involved found this unsettling and felt it affected the interaction. It raises questions about transparency, consent, trust and introducing technology into spaces where psychological safety is essential.
Another example is recruitment, where some candidates rely on AI-generated responses during interviews.
When hiring people for roles within the mental health space, you are looking for evidence of someone’s experience, judgement and ability to reflect. The concern is not the use of technology itself, but whether we are genuinely hearing from the individual. In mental health roles, understanding how someone thinks, reflects, responds and relates to others matters far more than an impressive-sounding answer.
These examples show that as AI becomes more embedded in everyday practice, we need to think carefully about where and how it is used.
Is AI always neutral?
Another important consideration is bias.
Large language models (LLMs) learn from vast amounts of information that inevitably reflect society’s assumptions, perspectives and inequalities. Its outputs are shaped by developers, training data and whose voices are represented or absent.
In mental health, this matters enormously. Our understanding of distress, wellbeing, relationships and help-seeking is influenced by culture, identity, circumstances and lived experience.
Could certain experiences be misunderstood or underrepresented? Could confident-sounding responses reinforce stereotypes or unhelpful assumptions? Could AI simply validate existing beliefs rather than encourage reflection?
These questions deserve ongoing attention, particularly when supporting people whose experiences may already be marginalised or misunderstood.
Curiosity rather than fear
I believe we need to embrace development with a curious mind.
We need to understand how AI is being used, including by the people we support, and create opportunities for learning, supervision and open discussion. Clear organisational policies, appropriate safeguards and the confidence to challenge potentially harmful uses are equally important.
AI could bring meaningful benefits to mental health services. Yet perhaps its greatest challenge is encouraging us to become clearer about what we value most. We should approach AI with curiosity rather than fear, and openness rather than unquestioning acceptance. Ultimately, technology should enhance our ability to support people, not replace the human connection at the heart of that support.
Miriam Mintz is Clinical Supervision Manager at Able Futures, overseeing clinical governance, supervising Mental Health Coaches, and ensuring high-quality, evidence-based support through the Access to Work Mental Health Support Service delivered by Department for Work and Pensions. She is passionate about improving workplace mental health, accessibility and inclusion, particularly for neurodivergent individuals, and contributes to wider policy discussions and recommendations to help people stay well in work linked to the Government’s Get Britain Working agenda.