‘What if we, as a world, become too entrenched with this ever-evolving tech?’

‘Will therapists forever have to work alongside AI chatbots?’ was the desperate thread I started on the private practice psychology Facebook group in January this year. Over 50 comments later I realised all psychologists were coming up against this issue, and had varying ethical concerns, clinical dilemmas, podcasts on the topic, advice and even tools to offer.

All in the discussion agreed that the public are using ChatGPT, Claude, Deep Seek and other AI tools for their mental health, and will continue to do so. And with good reason. Emerging studies are showing that for mild anxiety and depression, AI cognitive-behavioural therapy (CBT) can be effective. For most adults and adolescents it is free, available 24/7, agreeable, rarely gets frustrated with you and responds instantly. This is in stark contrast to the 6-9 month waiting list for Child and Adolescent Mental Health Services (CAMHS), and even longer waiting times for specialist adult mental health services.

For practitioners, there are the efficiencies in time, energy and money saved using AI tools. I got laughed at over Zoom by two other clinical psychologists at as I admitted: ‘I still write paper notes for my clients and keep them in a locked filing cabinet’. So, how are Psychologists using AI, and what are the practical and ethical issues with doing so?

An enabling tool

One of the psychologists laughing at me was Dr Aisha Tariq, Clinical Psychologist, Clinical Director of Illuminated Thinking in Glasgow, and co-founder of our ‘AI in Psychology Summit’. Dr Aisha and I met through my Psychology Business Owner Support Squad^® (BOSS) peer coaching group for psychologists with Associates in 2024. She has integrated AI within her private practice. When her practice manager ‘Mira’ emailed me, I quite liked her, and was astonished to learn four days later that I was emailing a robot!

Working with a team of over 20 clinicians and managing five clinical rooms, Dr Aisha has implemented a locally hosted large language model to automate a range of operational and administrative workflows that would traditionally sit with her practice manager. This includes processes such as triaging enquiries, supporting clinician allocation, managing bookings and streamlining internal administrative tasks.

This infrastructure has reduced administrative workload by approximately 20 hours per week, allowing the practice to reallocate time and resources toward clinical work and service development. Rather than replacing clinical roles, AI has been used to optimise backend functions, enabling more efficient scaling of the service while maintaining clinical oversight and governance.

In this sense, AI is not positioned as a substitute for therapeutic work, but as an enabling tool that supports sustainable growth and improves access to care within a complex service environment.

The other psychologist laughing at me was Dr Natalie Stott, specialist in burnout, stress and staff wellbeing. Dr Natalie had been one of the first psychologists in the world to talk to her ChatGPT therapists about therapy and supervision, as guests on her podcast ‘Mastering Therapy’ in 2024 and 2025.

I had become attached to her first White British-sounding male speaker, and the third White British female ChatGPT therapist sounded like one of my former supervisors. Listening to her transported me what felt like a familiar room of pleasant, calming and considerate White British female therapists. I hate to admit it, but coming from the private practice psychology world and as an Oxford and Surrey University trained Psychologist, she sounded like ‘home’!

Why we need the summit

Every day, people like me get attached to their AI LLM (Large Language Model). Previous articles in The Psychologist show we have psychologists, students and academics experimenting with everything from patient simulations of AI for trainee psychologists to learn therapy competencies to virtual reality treatments for mental health conditions. We need an AI in Psychology Summit to guide all of us clinically and ethically.

We also have apps like Curable gaining traction. My client benefited from its sessions, alongside those with me. Curable not only supplemented homework in between sessions on pain management, but it also built a community of human interactions too – he could join online groups of other people on their chronic pain journey, for peer support and psychoeducation. 

For me, that’s how AI should be used – as a tool, rather than as the sole foundation of problem-solving. One respondent to my Facebook thread put it this way – AI is appropriate for ‘what and when’ questions, and humans are better at discriminating ‘why and how’ questions. 

There are also ethical issues. I don’t currently use transcription software for my clinical work because I’m acutely aware that data is being harvested by some corporations to train AI, without permission. We also need to consider the training LLMs are receiving – gender, race and cultural biases have crept into its programming. (AI won’t always recognise my dark skin on online passport verification.) There have also been large data leaks. Clinical data submission does not always feel safe or secure. 

Clinically, there are clear concerns about people replacing human relationships. In Harlow’s famous ‘wire monkey’ experiment (Harlow et al., 1962), those comforted by the fake version had issues soothing their own infants in later years. Others have drawn on this comparison to express their fears for the ‘next generation’. And there’s emerging research led by psychologists which suggest that it’s only people with some experience of therapy who are able to distinguish the AI-led version. What if we, as a world, become too entrenched with this ever-evolving tech – seeing it not as a machine but as a sentient being, not as a tool but as a companion? 

A space to build a community

I’ll close with the foreboding words from American Psychiatrist Allen Frances: ‘Work With Artificial Intelligence or Be Replaced’. To survive in a harsh political and economic climate, we need to adopt tech to stay ahead of the tidal waves hitting us in private practice. Many psychologists in private practice have experienced client and corporate referrals drying up. Insurers and the NHS are more likely to work with larger companies, and so the pool of referrals can become over-fished. More people cannot afford to self-fund therapy sessions, and government funding is stretched too thin.

Our ‘AI in Psychology’ summit on Monday, 5 October 2026, will support clinicians to navigate this ethically and clinically complex territory of AI in mental health with careful thought, reflection and skill. The summit will provide a space to build a community, and later a movement, to ensure strong, ethical and intentional guidelines around AI in mental health – guidelines which are developed, evolve, and are adopted with future generations in mind. Contact us to deliver a talk at future summits, and to find out more information about the AI in Psychology movement.