AI in mental health support: a valuable tool or a dangerous force

Artificial intelligence is rapidly transforming the way we live, work, and connect. As we lean on GPTs more and more to answer every conceivable question, there’s a temptation to use them to offer guidance, support and even diagnose nuanced issues such as mental health concerns.

While the breadth of AI’s knowledge is unquestionable, is it still necessary for us to exercise some caution around gathering all our answers from Doctor AI?

Venture Zero Mental Health Lead, Clare Blunt, takes a closer look at the pros and cons of this complex issue.

The potential: accessibility and innovation

Mental health services around the world are under immense pressure. In the UK alone, NHS waiting lists for psychological therapy often stretch for months, and private sessions can be prohibitively expensive. With this in mind, it’s hardly surprising that people are turning to AI-powered tools that are always available – providing instant support when someone needs it most.

There are also several dedicated apps now on the market that have been developed in consultation with healthcare professionals to specifically support those in need of assistance with their mental wellbeing.

Apps such as Wysa and Woebot use conversational AI to help users manage anxiety, depression, and stress through evidence-based techniques like cognitive behavioural therapy (CBT). These digital companions can offer coping strategies, encourage reflection, and track progress over time.

For individuals who might otherwise suffer in silence, there is growing evidence that AI tools can be of genuine benefit. They are non-judgmental, accessible 24/7, and often free or low-cost – a compelling combination when traditional services are stretched thin.

The risks: accuracy, empathy, and ethics

However, the same qualities that make AI appealing also highlight its biggest weaknesses. While chatbots can simulate empathy, they don’t feel it. A comforting algorithm can’t truly understand the nuances of human emotion or provide the reassurance that comes from genuine human connection.

AI’s inability or unwillingness to challenge users has also been seen as a negative. Typically, AI responds to first-person user inputs and can, in some instances, be guilty of a confirmation bias. Much of the value provided by a trained professional is to be an objective third party able to help individuals struggling with their wellbeing to think differently or approach their issues from a different perspective.

There are also serious concerns about safety and accuracy. A 2023 study found that some AI chatbots offered inappropriate or even harmful advice when users expressed suicidal thoughts. Without proper regulation or human oversight, the potential for harm is significant.

Given that AI learns from the information available across the internet, it is innately susceptible to using untrustworthy sources. After all, you can’t believe everything you read. The problem with AI is that no matter where it has sourced information from, it sounds pragmatic, reliable and convincing.

In August, the British Association for Counselling and Psychotherapy published a statement outlining how vulnerable people could be “sliding into a dangerous abyss” through repeated use of AI chatbots that cultivated “emotional dependence, exacerbated anxiety symptoms and self-diagnosis”. 

The consensus among mental health professionals is broadly one of cautious optimism about the use of AI in controlled circumstances but that without safeguards, digital tools could be dangerous.

Privacy is another major issue. AI systems rely on large amounts of personal data to function effectively – but where does that data go, and who has access to it? In a sector as sensitive as mental health, a single data breach could have devastating consequences.

Striking the balance: AI as a supplement, not a substitute

The key may lie in viewing AI as a supplement, not a substitute, for traditional care. Used responsibly, AI can enhance or complement human-led therapy – helping clinicians monitor patient progress, flag early signs of crisis, and improve treatment outcomes through data-driven insights.

For example, some NHS trusts are trialling AI tools that analyse patient notes and communication patterns to detect potential mental health risks before they escalate. When combined with professional judgement and human empathy, these systems could save lives.

Perhaps the question then isn’t whether AI is a valuable tool or a dangerous force but whether we’re wise enough to use it responsibly. For now, at least, the recommendation is that AI tools should be used in tandem with professional support and guidance.

Want to ensure that your business is equipped to identify mental health issues in the workplace and provide meaningful support where necessary? Let’s talk