Fourteen years ago I co-designed a mental health app with young people. Reading it again as a Clinical Safety Officer, I found seven hazards, and none needed a coding error to cause harm.
Clinical safety work in the NHS still lives in spreadsheets, Word templates and inboxes. Vikela is our answer for the organisations that deploy digital health and clinical AI. Here is what it does, who it is for, and the line we have drawn to keep our assurance work independent.
DCB0129 and DCB0160 in plain English - what each standard covers, who owns which one, the artefacts you'll be asked for, and how a manufacturer's safety case feeds a deploying organisation's own risk assessment.
A new research agenda from 18 experts names six knowledge gaps in AI diagnostic safety - and every one of them sits upstream of a UK hazard log.
DCB0129 compliance is not the same as clinical safety. Here is why the gap matters - and what good looks like.
A concept prototype for a UK Healthcare AI Safety Registry - and an open call for challenge from anyone working in patient safety, clinical safety or informatics.
HSSIB has launched a national investigation into ambient voice technology in hospitals. The timeline means boards deploying AVT this year won't have the evidence before they need the answer.
My piece in Nursing Times argues nurses should be in the room when a decision is made about which AI joins the team - not handed the result and asked to make it work.
A Values-Anchored Framework for Trustworthy AI Conduct in Practice - why sustainable AI governance starts with the values NHS organisations have already co-produced.