My name is James and I’m going to a hackathon (in the UK) in 1 month. Our team’s project idea is to use a custom GPT chat bot, along with our own ML image analyser, to help diagnose the top 22 types of rashes. I was wondering how realistic this project idea sounds, and if creating your own GPT can even be used in this way. Specifically, i was wondering if you can:
Take a GPT and put it into your own website’s chat bot.
Give it information about an image, take user input and spit out a result based on some training data, along with a confidence rating (so that we can discard all results under a certain threshold).
An AI image classifier tuned on labeled data seems like a possibility and a good area of research. It would take a large dataset of imagery and computation to train an open computer vision model.
GPTs, though, are AI agents that only operate within ChatGPT plus for subscribers. It is not an externally accessible product like API AI models from OpenAI. GPT-4-Vision also is not trained for this type of medical diagnosis.
It would be possible to make a GPT that has actions that interface with an API where the external trained vision model is located, but there wouldn’t really be much value added by the AI’s interpretation of vision results.
You’re completely correct. I believe i was in over my head with this project idea. We will stick to using a pre defined response chat bot. Thank you for helping me realise that. Have a good day!
Hello James, I am interested in which hackathon you are attending. If it’s PhaseOne, I’ll also be there and would be interested in speaking with you in advance. All the best. Michael