The new independent group will evaluate artificial intelligence tools used in dermatology, aiming to generate real-world evidence on safety, effectiveness, and clinical performance before wider NHS adoption.
A new UK consortium has been established to independently evaluate artificial intelligence technologies in dermatology, with its clinical leaders saying the initiative could help inform AI assurance frameworks internationally as health systems, including Australia’s, grapple with the safe adoption of diagnostic AI.
The British Association of Dermatologists (BAD) launched the Skin and Responsible Artificial Intelligence (SkinRAI) Consortium at its 106th Annual Meeting in Manchester.
The clinically led initiative will assess whether AI tools used in dermatology are safe, effective, equitable, and clinically useful before they are implemented more widely across the NHS.
The launch follows recommendations from the National Commission into the Regulation of AI in Healthcare, which called for trusted datasets, real-world evidence, ongoing evaluation, and post-market monitoring to underpin safe AI adoption.
Dermatology is among the specialties leading AI implementation, particularly for assessing and triaging skin lesions.
While the technology has the potential to improve diagnostics, streamline referrals, and increase capacity in high-demand areas such as skin cancer, experts say there is still insufficient evidence demonstrating that AI systems perform consistently across different patient groups and healthcare settings.
Dr Rubeta Matin, clinical director of the SkinRAI Consortium and chair of the BAD AI Working Party Group, told Dermatology Republic the consortium was still in its early stages and would initially focus on building the foundations for a credible UK-wide approach to AI assurance.
“SkinRAI’s role is not to promote AI, but to help clinicians, health services, and other decision-makers understand whether AI tools are safe, effective, fair, and useful in real-world dermatology settings,” she said.
Dr Matin said there was incredible innovation happening in the UK, producing AI tools that could achieve impressive results. However duplicating AI between hospitals and different patient groups was not that simple.
“Diagnostic AI is developing quickly, but the evidence base is still immature in important areas,” she said.
“Performance in one patient group or healthcare setting cannot automatically be assumed to translate safely into another.
“That is why SkinRAI’s emphasis is on evidence-led assurance, real-world evaluation and clinically grounded decision-making.”
The challenges being addressed by the consortium extended well beyond the UK, Dr Matin told DR.
“Although SkinRAI is UK-led, the challenges it is trying to address are international,” she said.
“Questions around evidence, safety, bias, clinical performance, data quality, and post-deployment monitoring are relevant to health systems around the world.
“We would expect parts of SkinRAI’s approach to be of interest overseas, and there are likely to be ways for researchers, clinicians, health services, and other contributors from outside the UK, including Australia, to contribute to, support or learn from the work.”
The consortium will bring together clinicians, researchers, patients, NHS organisations, regulators, and industry to develop shared standards, evaluation methods and practical guidance for assessing dermatology AI.
A key priority will be creating a national dermatology AI data asset that enables technologies to be evaluated using representative patient populations, clinical pathways, and healthcare settings.
“AI is already being used in some dermatology pathways in the UK, particularly for assessing and triaging skin lesions,” said Dr Matin.
“Over time, there may be opportunities for AI to improve diagnostics, support referral management, and help increase capacity, particularly in areas such as skin cancer.
“However, wherever AI is used, it needs to demonstrably improve patient care, support clinicians, reduce pressure on services, improve equity, or make pathways safer and more efficient.”
According to Dr Matin, independent assurance will remain essential even after AI systems are introduced into routine practice.
Related
She added that AI performance could change over time and vary according to patient demographics, skin tones, imaging sources and clinical workflows, making post-deployment monitoring a critical component of safe implementation.
“The SkinRAI Consortium has the potential to help shape international expectations for how dermatology AI should be evaluated,” she said.
“Its value would lie in developing robust, transparent and practical methods that others could learn from, adapt and improve, rather than in promoting a single rigid model.
“Monitoring also needs to continue after AI tools are deployed. Performance may change over time and may vary across patient groups, skin tones, image sources, workflows, and local services.
“The broad SkinRAI model of independent assurance, representative evidence, clinical leadership, and post-deployment monitoring will therefore have significant relevance beyond the UK.”
BAD President Dr Tamara Griffiths said the consortium would help answer fundamental questions about whether AI tools improve patient care, support clinicians, reduce service pressures, and deliver value before they are adopted more broadly.
The consortium is now inviting expressions of interest from clinicians, researchers, patient representatives, regulators, and industry partners interested in contributing to its work, with developers hoping its evidence-based framework will become a practical model that other countries can adapt rather than a prescriptive international standard.
