How the BMJ Group is helping AI companies stay up to date
Many AI companies are training large language models to reflect best practice, and some seek up to date evidence to do so. Healthcare IT Today interviewed Derrick Leung,
business development director at BMJ Group, at the HIMSS Global Health Conference and Exhibition in March, about how BMJ Group is working with software companies to bring
evidence-based content directly to clinicians. Read the interview.
Your questions on AI adoption
Healthcare professionals have questions about using AI in practice. To help, we hosted a webinar on 30 April with Keith Grimes, founder of Curistica, a former GP, and member of
the BMJ Future Health advisory board, and Helen Surana, co-editor in chief of BMJ Innovations and chair of the BMJ Future Health advisory panel.
More than 45 community members from 10 countries joined the discussion.
Grimes offered a starting point for any clinician approaching a new AI tool: "Understand what the thing you're using does, how it goes wrong, what to do if it goes wrong, and who
to tell."
- Questions from the community included:
A general practitioner in Ireland asked: "What specific weaknesses in general practice are most at risk of being amplified [by AI]?"
A student in Denmark asked: "What are the key barriers around safety and workflow integration, and what practical strategies can help bridge this?"
Watch the full recording (60 minutes) for the answers (question one at 48:58 and question two at 42:25).
Want to go deeper? Scroll down for our reading list on AI in healthcare. We've unlocked the paywalled articles for BMJ Future Health readers until 29 May.
You can also submit questions for our June webinar. Look out for this after the Community Spotlight interview.
Competing interests:
- Keith Grimes is director of Curistica, which provides compliance services to various health tech firms. He is a shareholder in Tortus.ai and holds honorary academic and clinical roles at the University of Warwick, UCL, and South Downs Health and Care Federation.
- Helen Surana is sole owner and director of Helen Surana Ltd and works as a freelance editor across several journals and other BMJ projects.
Stay informed in a rapidly evolving landscape
We have selected these papers from BMJ Group journals to help you keep pace with technological developments in healthcare.Research- Substantial amount of medical information provided by popular chatbots inaccurate and incomplete
- Acceptable accuracy for medical AI: a survey of physicians and the general population in Sweden
- AI model detects very early normally "invisible" tissue changes of pancreatic cancer (featured in Bloomberg News)
- New brain stimulation technique improves short term social skills in children with autism
- Clinician in the loop: a flawed solution for AI oversight | The BMJ [Unlocked for BMJ Future Health newsletter readers until 29 May]
- Helen Salisbury: AI medical chatbots—more hype than help
- Why are humans still in the loop with advancing AI capabilities?
- Innovation in Alzheimer's disease needs more than pharmaceutical solutions [Unlocked for BMJ Future Health newsletter readers until 29 May]
Community spotlight
The epidemiologist teaching healthcare professionals to question AI
"Every single number we deal with on the screen in an emergency context, those are people with their stories. They are mothers, children, fathers, grandparents," Mona Elbabarytold BMJ Future Health.
Elbabary, an epidemiologist with the World Health Emergency Programme at WHO's regional office for the Eastern Mediterranean, runs an AI literacy initiative for professionals
working with displaced populations and families living under war. According to Elbabary, the initiative responds to a global concern: AI deployment is moving faster than the ability
to use it safely, with many trusting the technology's output simply because it sounds convincing.
Born in Cairo and bilingual in English and Arabic, Elbabary returned to Egypt three years ago after 20 years in Sydney, swapping Australia's step counting and healthy eating
culture for Egypt's love of sweet treats, which she enjoys equally.
She arrived confident she understood the region, but quickly learnt how diverse it is, from the wealthy Gulf states to conflict affected Sudan and Yemen. The region covers 22
countries and around 700 million people, including the settings of some of the world's most complex humanitarian crises, among them Gaza and Afghanistan.
AI literacy, Elbabary says, is not a technical course on AI tools but the practical knowledge needed to use them safely. "It's about what AI can and can't do, how to use it
responsibly, and the risks and limitations built into this technology."
Elbabary said that the stakes of low AI literacy are higher in the region than in most other settings, because AI tools trained on data from wealthy health systems, then deployed in
fragile ones risk amplifying the inequalities they are meant to address. The cost, she said, is borne by people who already have the least.
To reduce that risk, Elbabary made four recommendations specifically for clinicians and emergency officers:
- Know where data goes and what can be shared
- Maintain a critical mind: a tool trained on high income country data may give a confident output in rural Africa, but is it relevant?
- Be aware of large language model (LLM) limitations, including hallucinations and bias
- Practise with prompting, since output depends on input quality
Clinicians and emergency officers need to remember that AI supports decisions but can never make them, she said, adding: "You don't want to add another layer of risk to people
going through the hardest hardship on the globe."
Elbabary spoke on this topic as part of our panel, "Digital health where it matters most: innovation in humanitarian response," at BMJ Future Health Middle East on 13 May. You can
purchase post-event access via the website.
Competing interests: Aside from working for WHO, Elbabary has no competing interests.
Question of the week from the community
A doctor in our community asked how to handle AI when there is no regulation yet. How are you handling this in your region? What AI software are you using in your hospital or
practice? Reply and let us know, and we'll cover it in our webinar in June: cschuster-bruce@bmj.com
Submit your work
If you’re working on a relevant project, we want to hear from you. Explore our calls for papers.- Geopolitics of global health (includes AI/digital health in the context of geopolitics)
- Doing ethics with AI (Deadline: 24 May 2026)
- Artificial intelligence in rheumatology (Deadline: 31 October 2026)
- Digital health and primary care (Deadline: 31 October 2026)
We also have a live call for "problems", which includes and opportunity to showcase your work at BMJ Future Health Europe in London.
Thank you for reading. We hope to catch you at one of our webinars or events.