Journal

Three UK clinical answer engines, three architectures of trust: direct source retrieval, national-plus-organisational search, and a curated GP-authored archive; six test questions ...

Single-condition guidelines collide in real patients; how ANPs can use AI to organise the collisions, seven sorting categories for a fourteen-medicine review, and why an AI-optimis...

A real citation can sit under a wrong answer: five citation failures, the 2026 evidence that even the best AI drug-information tools score below pharmacists, and the verification s...

Students use AI weekly while faculty confidence and policy vary wildly; why educator development, not student training, is the binding constraint, and the five-component faculty cu...

Institution-provided AI solves equity, privacy and governance problems ad hoc adoption cannot, and creates an accuracy-halo risk; the portfolio model, and the questions that decide...

When an AI tutor teaches the wrong country's guideline, responsibility distributes across student, platform, school and model provider; the worked case, each party's duties, and wh...

A published benchmark for AI patients: partial recall, emotional variability, meaningful ambiguity, question-contingent disclosure, consistent hidden state, safeguarding behaviour ...

Distributed programmes lack tutors, OSCE partners and specialist case exposure by geography, not by talent; where virtual patients and asynchronous tutoring genuinely narrow the ga...

MSC and HDR UK recommend graduates arrive with informatics, AI and data-governance capability; what that means in FY1 terms, scenario by scenario, and the difference between using ...

The AMC's Digital Capability Framework spans the whole continuum of education and practice in Australia and New Zealand; translated into concrete student capabilities, year by year...

Canadian students recognise inaccuracy and bias risks at high rates and three-quarters want formal AI teaching; what schools should teach beyond prompting, and where platforms end ...

Making assessment technically impossible for AI is an arms race medicine will lose; verifying independent capability is a design problem medicine can win, and the redesign menu alr...