©
Shutterstock
Dr Andrey Kormilitzin and Dr Graham Blackman both use Artificial Intelligence (AI) as part of their work into brain and mental health in the Department of Psychiatry. Recently they contributed to a government POSTnote (Parliamentary Office for Science and Technology) on AI and mental health care. POSTnotes are flagship informational reports to give impartial information to decision-makers on emerging research areas.
Dr Kormilitzin is Senior Research Scientist in translational artificial intelligence, while Dr Blackman is an NIHR Clinical Lecturer specialising in neuropsychiatric disorders, particularly psychosis.
Tell us how you are using AI in your research at the moment?
Dr Kormilizin: "As a technologist, am passionate about how advanced computational tools may be meaningfully helpful to clinicians, researchers and patients. In close collaboration with clinical colleagues at the Department of Psychiatry, I have been developing a range of AI/ML models across various areas. For example, a new digital triage tool, using a large collection of free text clinical notes and Natural Language Processing (NLP) methods, to alleviate the problem of “hidden waiting list” and streamline the referral process in adult specialist mental health services. More recently, my research has shifted to leveraging the semantic reasoning capabilities of privacy-preserving Large Language Models to a range of problems, including: self-harm recognition in patients with severe mental illness; identification of patients who meet criteria for difficult-to-treat depression, evaluating the response and treatments provided to children and young people with anxiety, and a new triage tool to streamline psychological assessment for young people with autism and ADHD."
Dr Blackman: "As a clinician-scientist, I am very interested in using artificial intelligence to develop clinical decision-aid tools for mental health care. Specifically, I am exploring AI’s potential to predict diagnosis and outcomes in patients experiencing psychosis, with the aim of informing clinical care. For example, I have been working on a clinical prediction model that applies NLP to electronic health records to identify patients presenting with psychosis who may have an underlying neurological, or other medical cause for their symptoms."
What are the benefits and implications for using AI in brain and mental health research and healthcare?
Dr Kormilizin: "Contemporary technologies offer great opportunities to automate repeated tasks and streamline many processes, such as enhanced early detection and diagnosis, effective triage, help develop personalised treatment plans, and enhanced recruitment for clinical trials. With all great benefits, there are critical aspects that should be considered, such as handling sensitive patient information, which requires robust data protection measures and strict regulatory oversight to prevent breaches and misuse. Also, ensuring that AI models are trained on high-quality and representative data to mitigate potential algorithmic biases."
Dr Blackman: "AI can integrate rich and diverse data sources to inform clinical decisions around diagnosis and treatment planning. While acknowledging the many challenges of AI research, one key benefit is its potential to drive precision psychiatry, shifting from a one-size-fits-all approach to personalized treatment plans based on an individual’s unique clinical and biological profile. Ultimately, the goal is to improve clinical outcomes for patients with mental disorders by ensuring they receive the right treatment at the right time."
What opportunities and challenges does it bring?
Dr Kormilizin: "AI has the potential to streamline data processing and diagnostics, enabling precision psychiatry by integrating data from wearables, neuroimaging, and genetic profiles to deliver personalised interventions and real-time monitoring. However, challenges include the “black box” nature of many AI algorithms that can undermine clinician trust and risks of algorithmic bias from unrepresentative training data, such as LGBTQI+ communities. Robust data governance, patient privacy, and adherence to ethical and regulatory standards are crucial. Additionally, integrating AI into traditional healthcare requires investments in infrastructure and clinician training for effective collaboration."
Dr Blackman: "AI offers the opportunity to rapidly process large volumes of information from various sources, such as electronic health records, clinical assessments, and neuroimaging data. This can accelerate decision-making and improve diagnostic accuracy. A key challenge is ensuring that AI is used effectively and responsibly within healthcare settings, which necessarily requires government involvement to establish appropriate governance and regulatory frameworks. Additionally, widespread adoption requires overcoming barriers related to training clinicians, addressing data privacy concerns, and integrating AI into existing electronic health systems."
What are the current policy implications for using AI in mental health care, and what do you think it is important for policy-makers to know about?
Dr Kormilizin: "Policymakers must ensure that AI’s integration into mental healthcare is underpinned by robust ethical and regulatory frameworks. This includes establishing clear standards for data privacy, transparency, and accountability, and investing in digital infrastructure to support standardised, high-quality data. Ultimately, decision‐makers should focus on evidence-based evaluations and public trust-building to ensure that AI tools enhance, rather than replace or even become detrimental to human-delivered care."
Dr Blackman: "A key area of focus is precision psychiatry, which leverages multiple data sources to make tailored predictions that can inform diagnosis and treatment. While ongoing research is essential in this area, the successful implementation of AI in mental health will require substantial investment, involvement from key stakeholders, upskilling the clinical workforce, and the building public trust."
How could the information you provided for the POSTnote make a difference and have an impact?
Dr Kormilizin: "Contributing to POSTnotes allowed me to share my hands-on experience with developing AI/ML tools and address the challenges of applying them to real-world mental health data. Given the sensitivity of mental health information, off-the-shelf AI/ML solutions must be adapted for ethical and clinically meaningful use. Successful deployment within psychiatry and across the NHS requires integrating stakeholder consultations, clinicians, technologists, and patients, from the project’s outset, ensuring that the tools are developed to meet the genuine needs of all end-users effectively and efficiently."
Dr Blackman: "As a clinician-scientist, the chance to contribute to a POSTnote was an opportunity to share my experiences both as a researcher using AI and as a clinician working in the NHS, as well as to discuss how advances in the field could one day be translated into clinical practice. Through the report, I hope policymakers are informed about the potential benefits of AI in improving mental health care delivery, while also highlighting the critical elements necessary for its successful implementation, such as the need for appropriate regulatory frameworks and engagement with key stakeholders."
Tell us a bit more about how you got involved in doing the POSTnote, what it involved, and would you encourage others to be involved in this and why?
Dr Kormilizin: "I was approached by one of the POSTnotes authors and was invited to participate in an online interview. During our discussion, we explored the promises and challenges of applying AI in mental health and psychiatry. The experience allowed me to reflect deeply on my own work as well as helped identify key areas that need further attention. I encourage my colleagues to engage in similar initiatives, as they offer an invaluable platform to bridge academic research with policymaking and ultimately improve clinical practice. Such collaborations facilitate interdisciplinary dialogue and ensure that emerging technologies address real-world needs in a meaningful way."
Dr Blackman: "I became involved as a contributor to the POSTnote after learning about the opportunity through a departmental notice. The process included a semi-structured interview covering key topics in the field. It has been a highly rewarding and stimulating experience, and I have significantly expanded my knowledge as a result. I strongly encourage other clinical academics to engage in similar initiatives. Working at the intersection of government policy and academia provides valuable insight into how research can inform real-world decision-making and contribute more broadly to society."