top of page
Search

AI for healthcare professionals: Why Clinicians Hesitate, and What Bright Med's "H.I." Model Offers Instead.

Healthcare professionals lose roughly two hours to admin and desk work for every hour spent with a patient — and most have had no practical, safe introduction to the AI tools that could claw some of that time back. That's the gap Bright Med's free webinar, From Hype to Help: Large Language Models in Practice, set out to close on Thursday 13 August 2026, led by Consultant Nurse and Bright Med founder Alberto Barea.


Illustration representing administrative time reclaimed by healthcare professionals
Illustration representing administrative time reclaimed by healthcare professionals

The admin burden behind the burnout

Administrative burden is a well-documented driver of clinician burnout, and the numbers explain why. A 2016 US time-and-motion study (Sinsky et al., Annals of Internal Medicine) found that physicians spent 49% of their working day on electronic health records and desk work, against just 27% in direct contact with patients — roughly two hours of paperwork for every hour of care. A 2025 scoping review of large language models in clinical documentation (Woo et al., International Journal of Nursing Studies) found they can cut documentation time by up to 40%, with the usual caveat: every draft still needs a human check.

That's the whole premise of the webinar: AI should remove time from the equation, not judgement. As Alberto put it on the day, you still bring the same brainpower to set up the task and check the output — you just get your output faster.



Dermatology healthcare professional exploring an AI assistant on a laptop
Dermatology healthcare professional exploring an AI assistant on a laptop

Why so many clinicians are still hesitant

Despite tools like ChatGPT, Microsoft Copilot, Claude and Gemini sitting on most desktops already, uptake among UK healthcare professionals remains cautious. An updated UK survey of GPs found only 25% reported using generative AI tools in clinical practice, and 95% said they had received no professional training in using them. Dermatology teams show a similar pattern: UK dermatologists surveyed on AI in skin cancer management were broadly supportive in principle, but consistently raised medicolegal liability and the risk of widening health inequalities, and favoured what researchers termed an "augmented intelligence" model that keeps a human firmly in the loop rather than a fully automated one.

A few things sit behind that hesitation, and none of them are irrational:

  • A dose of ordinary technophobia, made worse by the fact that learning a new tool feels like one more thing on an already overloaded plate.

  • The "black box" effect — not knowing how an LLM actually arrives at an answer.

  • Fear of hallucinations — confident, fluent, entirely wrong output. Newer frontier models hallucinate far less than earlier ones (around 4–6% on general tasks, down from 15–50%+), but studies of medical case summaries still found hallucination rates as high as 64% when prompts weren't carefully structured.

  • Job security worries — a live and reasonable question, even if current evidence points to augmentation of admin tasks rather than replacement of clinical roles.



Two clinicians reviewing an AI-generated draft together, representing human oversight of AI output
Two clinicians reviewing an AI-generated draft together, representing human oversight of AI output

Introducing HI: Bright Med's answer to the trust gap

Bright Med's response to that hesitation isn't to push AI harder. It's to change what sits at the centre of the product. The tools Bright Med is building for its members won't be branded as AI — they'll be HI: Hybrid, Human Intelligence.

Every tool is built on content created by human dermatology specialists first. Artificial intelligence technology is used to make that content easy to access and use — searchable, structured, quick to apply at the point of care — but it never replaces the clinical authorship behind it. The content stays human, written by specialists for specialists, with AI doing the job of making it usable rather than the job of thinking for you.


What attendees said

"Thank you for organising your talk on AI. I am very new to this area, but you explained everything very clearly and made it very easy to understand. In particular, I liked how you explained large language models, the different types, and the key points to consider when using them effectively. I look forward to attending further talks on AI." — S.S., webinar attendee
"Thanks Alberto, that was really good. I especially liked your equation about the tasks where AI can be time-saving." — Dave, webinar attendee

Key takeaways: AI for healthcare professionals

  • Admin burden is a measurable, evidenced driver of burnout.

  • Hesitation towards AI among healthcare professionals, dermatology included, is grounded in real concerns: trust, hallucination risk, liability and workload — not simple reluctance to change.

  • Structured prompting and newer models both meaningfully cut hallucination risk, but a human still checks and signs every output.

  • Bright Med's upcoming HI tools keep human expertise at the centre, using AI to make that expertise easier to reach — not to replace it.

This was a free webinar, open to Bright Med's community of UK healthcare professionals. If you missed it, keep an eye on the Bright Med events page for the next session and for updates on the HI tools as they launch.


Frequently asked questions

What is a large language model (LLM)? An LLM is a program trained to predict the most likely next word in a sequence, based on patterns learned from vast amounts of text. It doesn't "know" facts the way a database does — it generates fluent, plausible language, which is why every output needs a human check before use.

Is it safe for healthcare professionals to use tools like ChatGPT or Copilot? It can be, for admin tasks, provided you never enter patient-identifiable information into a public tool, use only organisation-approved software with a completed Data Protection Impact Assessment, and treat every output as an unverified first draft that you check before use.

What is Bright Med's HI model? HI stands for Hybrid, Human Intelligence. It's Bright Med's approach to its upcoming member tools: content written and verified by human dermatology specialists, made easy to use through AI, with the clinical expertise — not the algorithm — at the core.


References

Sinsky C, Colligan L, Li L, et al. Allocation of Physician Time in Ambulatory Practice. Ann Intern Med. 2016. https://pubmed.ncbi.nlm.nih.gov/27595430/


bottom of page