The Promise of AI and Potential for Scaling

The potential benefits of artificial intelligence (AI) for supporting under-resourced mental health services is considerable. These emerging technologies could improve access to services, streamline service delivery, and support clinical decision-making. Those of us working in clinical services can see how AI-supported tools may extend our capacity to support underserved populations while alleviating pressure on existing and overstretched systems.

However, it is essential to approach any change with measured realism. While AI may enhance certain aspects of care, at this time, its capacity to safely and reliably replicate the depth and nuance of human therapeutic relationships remains to be determined. The value proposition of AI in mental health may ultimately be more modest than anticipated, particularly in domains requiring contextual understanding and complex clinical judgment. But in addition, the current state of mental health systems means that AI tools will struggle to perform important tasks such as helping consumers identify and navigate non-digital pathways and services and other important tasks which require therapists to facilitate referrals between non-integrated services and systems.

Moreover, some of AI’s perceived strengths — such as accessibility and immediacy — may carry unintended risks. Over-reliance on automated tools could lead to increased dependency on digital interventions and delayed help-seeking in circumstances where more detailed clinical assessments are warranted. These risks must be carefully monitored and mitigated through thoughtful design and robust ethical and governance oversight.

For these and other reasons, the mental health workforce must be engaged and supported. This can be achieved by actively involving clinicians and service providers in the design, implementation, and evaluation of AI tools. Their expertise is vital not only for ensuring safety and efficacy but also for fostering trust and adoption. In recent months I have been introduced to several new AI chatbots, none of which were developed within a clinical service, and none of which could be applied beyond a very limited scope of practice. From those experiences it clear that the successful adoption of AI tools will require investment in collaborative innovation, training, and change management.

Addressing Ethical Concerns and Clinical Safety

While AI presents exciting opportunities for mental health care, its integration must be guided by evidence, engagement with the mental health workforce, and a clear understanding of its limitations. The future of mental health services will depend not only on technological advancement but on our ability to centre human expertise and compassion without compromising high standards of clinical safety or care.