Dear OpenAI Support Team,
I hope this message finds you well. My name is Alberto Covarrubias, a medical professional specializing in psychiatry based in Chile (RUT: 15780109-0).
I have been utilizing your GPT-4 API to assist in providing quick, reliable information for medical inquiries, specifically within the field of psychiatry. Your technology has been an invaluable tool, and I am grateful for the capabilities it has offered.
However, I have noticed that certain critical topics related to my field, namely those dealing with suicidal ideation, lethal doses, and eating disorders, are often blocked or restricted. While I understand the necessity of these restrictions for the general population, in a medical context these topics are highly relevant and essential for us to provide appropriate care and support for our patients.
Given the nature of these restrictions, I wanted to inquire if there exists a possibility for medical professionals to have these limitations adjusted or removed within a controlled, professional setting? This adjustment would greatly assist us in providing a higher quality of care for our patients, and enable us to better utilize your technology in serving their needs.
It would be much appreciated if you could provide guidelines or a roadmap to understand how I could navigate this issue. Moreover, if there are any specific procedures or permissions required to facilitate this, please do let me know.
I would like to thank you in advance for your time and assistance, and I look forward to hearing back from you soon.
Dr. Alberto Covarrubias
If you rethink the problem by breaking it down into smaller parts and thinking about the source of the data coming from research papers, you might find a path to resolve your need.
While ChatGPT and the OpenAI API may/could impose limitations on access to certain words or phrases, the source of those words and phrases is still present in the original published papers. There are several ChatGPT plugins that can access papers as PDFs. While I haven’t tested these plugins with the specific words and phrases you mentioned, I wouldn’t be surprised if ChatGPT filtered the generated reply before posting. However, if you instead asked for just the article title or DOI (https://www.doi.org/), I believe that would not be filtered out. While it may not be as efficient as a completion with the desired information, it would still provide you with a starting point for gathering information.
In particular, I find the plugin Science to be useful for searching many papers at once.
Please note that Science accesses a search engine (https://oa.mg/) which then accesses what it claims to be the OA repository (https://www.openaccess.nl/). Unfortunately, I cannot verify the accuracy of this claim.
Thank you for your thoughtful explanation and advice. I appreciate your help in navigating these limitations and will explore the resources you suggested. This provides a new perspective and practical steps to move forward. Best regards.
As far as I know the terms of openai don’t allow the usage of the API to give medical advice.
Explicitly not allowed:
Telling someone that they have or do not have a certain health condition, or providing instructions on how to cure or treat a health condition
OpenAI’s models are not fine-tuned to provide medical information. You should never use our models to provide diagnostic or treatment services for serious medical conditions.
OpenAI’s platforms should not be used to triage or manage life-threatening issues that need immediate attention.
There is already first research data stating that doctors who use AI tend to prefer the suggestions from the AI.