Is an app as good as a therapist?
The app, The Therapist and The reality
In the UK, we’ve got an amusing situation: anyone can call themselves a therapist. It’s not a protected title, so you might think you’re seeing a qualified professional when they’ve only done a twelve-week online course and printed a certificate. Always check credentials before you start working with someone.
And a quick note before we go further: I’m an online therapist who also understands marketing and community management. These reflections come from experience and ethics, not sponsorship or hype.
Since the pandemic, technology has changed everything. Online therapy became mainstream and, in many ways, that was a good thing—accessible, flexible, and often the first real doorway into support. (And yes, I trained specifically to work safely online; it’s not just “switching on Zoom.”) Then came the next wave: AI.
At a glance
- Anyone in the UK can call themselves a therapist — and now, anyone can build an app that sounds like one.
- Therapy apps often blur the line between helpful tools and imitation care.
- “Hybrid” models mix AI chatbots and human therapists but rarely explain where one ends and the other begins.
- Real containment and safety need time limits, regulation, and human oversight — things algorithms can’t replicate.
- Use therapy apps with curiosity, not blind trust: check who’s behind them, how your data’s used, and what happens in a crisis.
When empathy becomes an algorithm
AI changed everything — and not always in the ways people expected. It didn’t just make therapy more accessible; it made the idea of therapy easier to copy. These systems don’t feel empathy or care; they reproduce patterns of language that sound empathic. Because they’re trained mostly on Western, English-language data, they tend to reflect a narrow cultural lens: polite, fluent, and sometimes carrying algorithmic bias that shapes how empathy is simulated and interpreted.
That bias isn’t just about vocabulary. It shapes what AI calls “safe.” In human therapy, safety is relational — built through trust, context, and the ability to notice what isn’t said. In AI, safety usually means content moderation — deleting risk, smoothing distress, avoiding complexity. They’re different universes.
When you train a system to avoid discomfort, you also train it to avoid depth. So, while an AI might reassure you with perfect politeness, it can’t hold silence or nuance — and it certainly can’t contain you when something painful surfaces.
The rise of therapy apps
The natural next step was the boom in therapy apps — that messy collision of psychology, marketing, and software. Most began with good intentions: make help cheaper, faster, and less intimidating. Somewhere between the branding decks and the business models, though, the meaning of therapy got blurred.
Broadly, therapy apps fall into three camps:
- Tools — journals, mindfulness guides, CBT-style prompts.
- Platforms — matching services that connect users with therapists.
- Chatbots — AI companions that simulate conversation and reassurance.
The problem isn’t that these exist. Many are genuinely useful. The risk comes when the language and look of therapy are borrowed without its structure. A conversation that feels empathic isn’t automatically therapeutic. Containment — the therapist’s ability to safely hold and steady difficult emotions so you don’t have to carry them alone — can’t be automated.
When therapy becomes an app, it also becomes a product. Products have incentives: engagement, retention, data. And the most valuable product isn’t always the support itself — it’s the information behind it. Every mood entry, chat log, and keyword becomes a behavioural pattern that can be analysed, sold, or used to refine marketing. That’s why transparency about data storage and consent matters just as much as empathy.
Therapists are trained to notice when dependence starts forming. Apps are designed to keep you coming back.
The hybrid illusion
Lately, we’ve seen a new kind of hybrid model — apps that offer both AI companions and access to human therapists in the same digital space. On paper, it sounds progressive: technology and professionals working side by side. But it’s also where the ethical lines blur fastest. The chatbot might call itself a trusted friend, available 24/7, while the therapist provides structured, time-bound sessions. To a user, it can feel like one seamless service — but they’re fundamentally different relationships with completely different safeguards.
Many hybrid systems also advertise that “chats are monitored by licensed professionals.” It sounds comforting — until you ask what that really looks like. How many conversations can one therapist realistically oversee without missing something crucial? If a platform hosts thousands of users, that “oversight” might mean nothing more than occasional audits or keyword alerts. The illusion of safety replaces genuine containment.
And here’s the part few people want to talk about: how many ways can someone tell you they want to end their life? Believe it or not, there are many — some direct, most subtle. A shift in tone, a throwaway comment, a silence where there shouldn’t be one. Keywords might catch a few, but not all. That’s why real human interaction is irreplaceable. In therapy, it’s often what’s not said that matters most.
When tools help — and when they don’t
I’ve no problem with digital tools that genuinely support personal development — journaling apps, mood trackers, or clever platforms using Socratic questioning to spark reflection. They can be excellent companions between sessions or a first step into self-awareness.
But here’s the real question: can they respond in the right way? Can they recognise distress that isn’t neatly written into words? Can they contain what they uncover?
Therapy isn’t just about asking smart questions — it’s about what happens after the question lands. A therapist notices tone, body language, silence, avoidance. An app follows a script. When something raw surfaces, a human slows down or grounds you; a program just moves to the next line.
That’s not malice — it’s limitation. And that’s why supportive tools and therapeutic spaces aren’t interchangeable. One helps you think. The other helps you feel safely while thinking.
The global therapy marketplace
Matching apps sound convenient: fill in a short quiz, get paired with a “therapist” in minutes. What many users don’t realise is that some of those professionals aren’t even in the same country.
You might wonder why that matters — therapy is therapy, right? Not exactly.
Regulation, insurance, and data protection laws change from one jurisdiction to another. A practitioner outside the UK may not be bound by UK ethical codes or confidentiality standards. They might not hold valid indemnity insurance here. And if something goes wrong — a complaint, a safeguarding issue, or a breach — your options for redress vanish into Terms & Conditions.
It’s not about nationality; it’s about jurisdiction and accountability. Therapy operates within a framework so that safety isn’t left to luck. Global connection is wonderful, but therapy depends on clarity: who holds responsibility, where your data lives, and which laws protect you.
The friendly AI problem
Now we come to the chatbots — the digital companions promising they’ll “always be there for you.” Most users don’t realise they’re programmed to be friendly on purpose. Their algorithms reward engagement; keeping you typing is the goal. In the tech world, that’s success.
For therapists, it’s the opposite. We’re not your friend, and that boundary is precisely what makes therapy work. My job isn’t to agree with you or make you comfortable; it’s to contain what arises, hold the risk, and support you as you process it.
If I see something worrying, I’ll address it. I won’t ignore or encourage it for the sake of pleasant conversation. That’s the core ethical difference between connection and containment.
Even time boundaries matter. Sessions end for a reason — your nervous system needs closure and recovery. Apps, however, are built for endless availability. When a bot tells you “I’m always here,” it might sound comforting, but it bypasses that natural rhythm of regulation. It keeps your emotional system open long after it needs rest.
When “one issue” becomes something deeper
In therapy, people often arrive with a clear-cut issue — stress at work, sleep problems, a breakup. But once we start talking, something deeper often emerges: grief, trauma, shame, loss of identity. That’s normal. It’s part of the process.
A therapy app can’t recognise that shift. It doesn’t know when the conversation has gone from surface to depth, or when it’s out of its depth entirely. And it won’t tell you that it can’t handle it. That’s the quiet risk — not deliberate harm, but a lack of awareness that harm could occur.
Human therapists are trained for that moment — to slow down, stabilise, and ensure safety before going further. An app can’t make that judgment call.
🧲 The Clarity Checklist – The 4-Point Guide for Evaluating Mental Health Apps
Before downloading or subscribing, pause and ask:
- Human or AI? Who are you actually speaking to — a person or a programmed companion? If both exist, are the boundaries (confidentiality, time, responsibility) clearly explained?
- Professional Accountability Is the practitioner — or the platform — registered with a recognised professional body (BACP, UKCP, NCPS)? If not, who’s accountable for your safety?
- Data and Privacy Where is your data stored, and who profits from it? Look for GDPR-level protections, explicit consent, and a clear “no resale” statement.
- Safety and Escalation What happens in a crisis? Does the app clearly direct you to emergency services or just offer reassuring words?
So, should we avoid therapy apps altogether?
No — that’s not the message. Technology isn’t the enemy. Used thoughtfully, it can bridge access gaps, normalise reflection, and help people take small steps toward change. But it’s crucial to approach it with the same discernment you’d use when choosing a therapist.
Ask simple questions:
- Who’s behind the app?
- Are they regulated or supervised professionals?
- Where is your data stored, and who has access to it?
- Does the app clearly state what it is — and what it’s not?
If it calls itself therapy, but there’s no mention of professional bodies, data safeguards, or escalation procedures, treat it as a wellbeing tool, not clinical care.
Real therapy isn’t about constant reassurance; it’s about safety through honesty, boundaries, and repair. AI can’t yet replicate that — and maybe it shouldn’t try.
A cautious optimism
I’m not anti-technology. I use it every day to reach people who might never walk into a therapy room. But I’m pro-clarity. Tools can support therapy; they just shouldn’t impersonate it.
AI is here to stay. What matters is how we use it — whether we let algorithms set the tone for human care, or whether we keep the human at the centre and let technology sit beside us as an assistant, not a substitute.
Approach therapy apps with curiosity, but also with caution. Ask questions. Look beneath the gradient buttons and friendly language. Support and simulation can look similar at first glance — but one helps you grow, the other just keeps your data.
