As someone who’s spent countless hours immersing myself in the world of artificial intelligence(AI) and its applications in mental healthcare, I wanted to write up my own thoughts on it. Firstly, worth saying I won’t pretend to be an expert in this, but I am someone who’s dedicated significant time to understanding this technology because I believe it represents the most significant shift in the last decade, much like the internet and mobile phones and whilst the change is scary to many, with every change becomes opportunities.
The reality is stark: AI isn’t coming to mental health–it’s already here. The question isn’t whether we should engage with it, but how quickly we can adapt to remain relevant and effective in our practice and our industry.
What Actually Is AI?
Personally, whenever I speak to any clinician in the industry, they say they use AI for things like learning more, “ I use chatgpt” is the most common response. For me though, I didn’t really understand what AI was for a long time, I just quoted ChatGPT and in all honestly that is what I thought AI was, which looking back was naïve of me.
So what is AI?
Artificial Intelligence is essentially software that can learn from information and make decisions or predictions based on patterns it discovers. Think of it like this: instead of following a rigid set of pre-programmed instructions, AI can adapt and improve its responses based on the data it processes, eg the data read or input manually.
Being ADHD I love analogy’s: Imagine you’ve been a therapist for 20 years. Over that time, you’ve developed an intuitive sense for recognising when certain clients might be at higher risk of relapse, or which therapeutic approaches tend to work best for specific presentations. You’re not consciously following a checklist-you’re drawing on thousands of interactions and outcomes to inform your clinical judgment.
AI works similarly, but instead of 20 years of experience with hundreds of clients, it can analyse patterns from millions of data points in seconds. For example, it might notice that clients who use certain language patterns, report specific sleep changes, or show higher engagement levels with assignments set by the therapy are more likely to experience symptom improvement with cognitive behavioural interventions versus other approaches.
The crucial difference: While you bring empathy, intuition, cultural understanding, and the ability to read between the lines, AI brings the capacity to spot statistical patterns across enormous datasets that no human could ever process.
“Ai will replace therapy” is something I hear a lot, initially I did think it will have a huge impact, but not as much as other industries. A recent Mckinsey report stated that only 35% of a therapists time spent is potentially automatable, meaning it is one of the lowest industries to be effected. AI isn’t a robot therapist trying to replace human connection, nothing can ever replace that, our industry is all about connection and that can’t be replicated by AI. It’s not reading minds or making mystical predictions. It’s sophisticated pattern recognition software that can process information at scale and speed that humans cannot match.
In practical terms for your work: AI might analyse your session notes (with proper consent and privacy protections- another discussion) and notice that clients who mention certain themes tend to benefit from specific interventions. Or it might flag early warning signs in routine questionnaire responses that suggest someone needs more intensive support, even before they’re fully aware of the change themselves.
The key insight is that AI doesn’t replace clinical judgment—it provides another layer of information to inform your decision-making, much like how we use assessment tools, outcome measures, or consultation with colleagues to enhance our understanding of our clients’ needs.
The Current State: More Advanced Than You Think
The evidence is overwhelming. AI-powered mental health tools have become a $2 billion market in 2025, growing at an extraordinary 34.3% annually. This isn’t speculative growth-it’s happening right now, driven by real clinical outcomes that demand our attention.
Recent clinical trials are producing results that should make everyone take notice. The groundbreaking Dartmouth study in 2025 showed that participants using AI therapy experienced a 51% average reduction in depression symptoms-results that mirror what we see in traditional therapy settings. A comprehensive meta-analysis of 18 randomized controlled trials involving 3,477 participants found consistent improvements across different AI platforms, with depression symptoms improving by an effect size of -0.26 and anxiety symptoms by -0.19.
Of course, it’s early research, but as we know with any research- early results with positive outcomes will mean more investment into the research and more adoption over time.
Does our industry use AI already?
43% of mental health professionals are already using AI, primarily for research and report writing- driving efficiency. Yet, the reason I am writing this is because many of us remain unaware of how rapidly this landscape is changing.
Why hasn’t it grown yet?
AI has been in our everyday conversations for at least a year now, with the huge investment going into AI we can see that it is here to stay, but why hasn’t it already taken over? My opinion on this that it is about trust. Like any huge change, for example mobile phones, we all remember thinking “ I’m never going to use a mobile, I’ll just go around peoples houses if I want to talk”, AI is very much there at the moment. The general public are still sceptical of AI because they don’t truly understand it, therefore implementing AI into our lives isn’t easy. However, what we are seeing is our mobiles, computers( Microsoft have co-pilot) are already getting us to utilise it, it’s slowly feeding into our everyday lives. The more this continues to happen, the more reliance we place on it, but also we trust me it more. When more people trust it, it is much easier to implement AI into industries, which is what is going to happen in 2026 and why it’s so important our industry understands it more, before it’s forced onto us.
So, you might be thinking, why should I care about AI? Will it actually affect me or my practice? To answer that, even if you don’t adopt AI into your practice, the general public are already using it, Google’s global search share dropped to 89% recently, the first sustained dip since a decade ago- all due to AI.
How AI Is Changing the Search for Therapists and Rehabs
AI is fundamentally transforming the way people find therapists and choose rehab facilities. Just a few years ago, searching for a mental health professional or addiction program meant wading through long lists online, looking at location or availability, hoping to find a good fit based on shortest descriptions and maybe a photo. Now, AI-powered platforms are radically improving that process for both clients and clinicians.
Modern platforms use AI algorithms that take each person’s preferences, symptoms, life circumstances, and sometimes even communication style into account. Instead of simple filters (“therapist near Glasgow who works with anxiety”), AI-driven systems analyse much richer datasets-things like brief in-app questionnaires, past therapy outcomes , even textual analysis of freeform descriptions. The algorithm then matches individuals with therapists or rehabs most likely to fit their specific needs and personalities.
A recent UK study found that 37% of adults searching for mental health support in 2025 used a digital platform with some form of automated matching. These algorithms are not just about efficiency-they’re directly addressing one of the main barriers in mental health: the challenge of finding the right therapist, not just any therapist. Research from the US and Europe in 2024 and 2025 shows that clients are significantly more likely to continue therapy beyond the third session when initially matched via an algorithm that considers personal values, availability, modality, and clinical expertise versus a random or self-directed search.
For addiction treatment, some rehabs are adopting AI-guided triage. These systems process intake information and use risk prediction models to suggest appropriate levels of care or flag urgent needs, helping clients avoid lengthy waiting lists for the wrong types of programmes. Digital health platforms now claim up to a 40% reduction in “therapy shopping” (switching providers multiple times before finding a match), based on both internal analytics and external studies published in recent digital health reviews.
From a clinician perspective, this evolution also means new patients who come through digital platforms are better informed, more motivated, and have clearer expectations-often because the AI system has already guided them toward therapists with approaches or expertise that align with their needs. We saw a similar shift during Covid, pre Covid -patients very much were guided by clinicians on where to go for inpatient or further support, post-covid patients were information heavy, they would already have ideas of where they wanted/thought they should go and took this to the clinician.
In summary, AI matching is increasing access, improving retention, and helping both clients and therapists make better use of time and resources. As these tools become more nuanced, they’re poised to become the new default in how people discover, evaluate, and connect with mental health care and the research from the last two years confirms this momentum is only growing.
Understand how it is changing patients finding clinicians is important, the clinician needs to adapt to this in order to be ‘found’. But you may be asking ‘ ok, so how we are found is changing, but it doesn’t alleviate the concerns I have around AI overall’.
Understandably, patient care and safety are crucial to what our industry delivers and AI at the moment feels like it will cross boundaries, create concerns and not protect the patients privacy. What is the research saying about what clinicians are concerned about?
Trust and Professional Scepticism: As discussed earlier, my anecdotal understanding is that our industry is still very sceptical about AI, research is also varying levels of scepticism about AI implementation, particularly when it comes to diagnostic or treatment decisions. I totally understand why, our industry should and is naturally cautious about anything that might undermine the clinician’s role.
Workload and Infrastructure Concerns: Studies consistently identify infrastructure issues, psychological barriers, and concerns about increasing workloads as primary obstacles to digital health adoption among many healthcare professionals. The fear isn’t just about learning new technology-it’s about whether it will genuinely reduce our burden or simply add to it.
Safety and Liability Worries: Many practitioners express concerns about patient safety and medical liability when prescribing digital therapeutics, mainly because most of us aren’t sufficiently trained to assess technical specifications like validity and usability. The responsibility ultimately rests with the practitioner, regardless of what the AI recommends.
The Confirmation Bias Challenge: Research reveals that we’re significantly more likely to accept AI recommendations that confirm our preliminary diagnoses, while rejecting those that don’t align with our initial assessments. This highlights how our existing expertise can both help and hinder effective AI integration.
The Humanisation Risk: NHS leaders have issued warnings about AI chatbots providing “harmful and dangerous” mental health advice, particularly among teenagers and young adults. These tools cannot account for body language, visual cues, or the nuanced understanding that comes from human interaction.
Privacy and Data Security: AI systems require vast amounts of sensitive patient data, raising critical questions about consent, data ownership, and security. The risk of algorithmic bias perpetuating existing inequalities is real and requires constant vigilance.
Professional Dehumanisation: There’s a genuine risk that over-reliance on AI could erode the therapeutic relationship that forms the foundation of effective mental health treatment. We must preserve the human connection that makes therapy therapeutic. AI overall will make us lazier, less creative as a society if we are not careful, however many industries like our own are all about connection, human connection, which can never be replaced.
Whilst AI grows, more technology is implemented, anything in our industry has to alleviate the above concerns, or work with the clinician to ensure it doesn’t occur.
Are clinicians already using AI? We saw earlier a stat about how 48% of mental health clinicians are already using AI, but is our industry adopting it on a macro level(larger organisations)?
The Applications Already in Practice
Clinical Decision Support: AI systems are analysing vast datasets to identify patterns invisible to human observation, helping predict treatment response and identify early warning signs of relapse. Machine learning models can predict response to antidepressants using clinical data and EEG signals, potentially eliminating ineffective medication trials.
Administrative Efficiency: The promise here is significant-NICE estimates that digital therapies could save over 5,000 therapist hours per 1,000 people for PTSD, 2,500 hours for social anxiety disorders, and 600 hours for depression.
Enhanced Diagnostic Accuracy: AI is supporting more precise diagnosis by integrating diverse data sources including electronic health records, neuroimaging scans, genetic information, and real-time behavioural data. This multi-modal/disciplinary approach addresses the inherent challenges in mental health diagnosis where symptoms often overlap between conditions.
24/7 Support Systems: AI-powered chatbots and digital therapeutics are providing around-the-clock support that complements rather than replaces human therapy. These tools handle routine check-ins and skill reinforcement, allowing you to focus on complex therapeutic work requiring human insight. A recent AI robot therapy offering highlighted that the most common time it was used was between 2-4am, when no clinician is available to respond.
Will AI replace clinicians?
It’s a question that is being said a lot and is also why many are concerned; will our industry be obsolete because of AI?
I have asked this question to myself a lot, hence why I was so interested in learning more about AI to understand it and its limitations. I came across a fascinating use of AI in neurosurgery.
The studies were using AI to provide the role of the neurosurgeon, could they do their job(bar the physical application)? The results showed very interesting outcomes, that AI systems are now able to detect subtle abnormalities in brain scans-such as tumours and haemorrhages-with comparable or superior accuracy to experienced neurosurgeons. For example, one study compared deep learning models to the diagnostic performance of neurosurgical specialists, finding that AI delivered higher sensitivity in detecting intracranial haemorrhages on CT scans than human residents and practitioners. These AI systems excel in identifying tiny changes or early signs that can sometimes be overlooked during manual readings, making them powerful screening tools for initial triage, especially in time-sensitive emergencies or settings without around-the-clock specialist coverage.
However, the research also highlights clear limitations. While AI can quickly flag potential concerns and support clinicians, it sometimes misclassifies benign abnormalities or generates false positives-emphasising the need for human oversight.
The Path Forward: Augmentation, Not Replacement
The consensus in the latest literature and my own understanding is that AI is set to become an invaluable diagnostic aide, capable of enhancing-rather than replacing-the expertise and intuition of skilled practitioners. Used together, these technologies ensure data is reviewed faster and more accurately, but final decisions and holistic patient care still rely on the unique judgment of trained medical professionals.
The most promising developments position AI as an augmentation tool rather than a replacement. Dr. Nicholas Jacobson from Dartmouth captures this perfectly: “I see AI helping provide mental health support to the huge number of people outside the in-person care system. Person-to-person and software-based therapy can work together”.
So where do I see AI changing our industry?
The future model I see emerging involves:
• Affordability- I see a future where AI will be available for those that can’t access private healthcare, whilst this area needs more work, it’s already being invested in.
• Therapists focusing on complex therapeutic work requiring human insight, empathy, and nuanced understanding.
• AI handling routine tasks like scheduling, compliance monitoring, and basic skill reinforcement.
• Assessment- it will be able to take a huge amount of data sets and provide the assessing team with more in-depth understanding, information on the patient.
• Care planning- AI will be able to develop care plans, using the best plan based on all the information available.
• Hybrid care models where AI provides 24/7 support between sessions while therapists maintain oversight and intervention capacity
• Predictive analytics helping you identify crisis risks before they escalate
Why Acting Now Matters
Every month we delay engagement with AI technology, we fall further behind in understanding how it can enhance our practice and serve our clients. The clinicians who thrive in 2026 and beyond will be those who started engaging with these tools in 2025-not as sceptics standing on the sidelines, but as thoughtful professionals learning to integrate AI ethically and effectively with the aim to enhance patient care, their practice and patient outcomes.
The transformation is already reshaping how you work, connect with clients, and understand the people you help. We can either lead this change or be swept along by it.
The choice is ours, but the window for proactive engagement is narrowing fast.
This is why I’m organising our conference on November 21st-not to push a technology agenda, but to create a space where we can explore these developments together, understand what it actually is, address our legitimate concerns, and prepare ourselves for a future where human expertise and artificial intelligence work in partnership to better serve those who need our help most.
Join me for “Navigating Change and Preparing for 2026 in the Mental Health and Addiction Industry” on Friday, November 21st. Tickets to be announced shortly- only 80 spaces available. Together, we’ll explore how to harness AI’s potential while preserving the human heart of therapeutic practice.
