Welcome to the first Emerging Health Tech Open Mic session for 2023, our fortnightly catch-up to chat about what innovative things are happening around both locally and internationally this week.
Working on something cool you’d like others to know about? Share it with us!
You will notice the new time and day (held fortnightly) for this year, hopefully allowing more people to attend. Please contact @jon_herries for your recurring invite, or click on the link below to jump into this week’s session.
Chat GPT
explain what the ehealth forum & HiNZ is
AI Bot ChatGPT Passes US Medical Licensing Exams Without Cramming – Unlike Students - https://www.medscape.com/viewarticle/987549
Creating personalised health content, changing the tone vs who the audience is
Deep cuts on philosophy and outcomes
Culture vs pathways? Survival? Something about photography? Tomatoes? Echolocating colour? (I got a bit lost while side questing here and don’t know what started the conversation)
Therapeutics product bill – what is a medicine, does this include food? Medical devices, does this cover fitbits to pacemakers?
What is the regulation that will come from this, and how do you manage such a diversity of products with a country our size?
Willow bark or your favourite chemist
Three groups Jon visited over Christmas/New Year
Matai – Gisborne – MRI research
Prostate Cancer
Pulse which appears to show inflammation of the brain
The quote the authors Our classifier is not fully reliable. In our evaluations on a “challenge set” of English texts, our classifier correctly identifies 26% of AI-written text (true positives) as “likely AI-written.”
Translation; toss a coin, and the odds of a true positive are the same or better than the app depending on how you read the ambiguity.
The authors wish to know if such products have utility. The answer is no they don’t and such nonsense does the whole field a disservice.
It is interesting though, as if you could put your thumb on the proverbial scales of horse racing odds or online advertising you stand to make a lot of money.
Unfortunately understanding the epidemiological consequences of poor accuracy aren’t really a profitable model (you spend a lot of profit checking that you didn’t miss something).
