ChatGPT - ethical in medical use? And more about DAN and aliens and the lying US government, hypothetically.

Eye roll

From a MedPageToday opinion article, Are AI Chatbots in Healthcare Ethical?

— Their use must require informed consent and independent review

by Benjamin Tolchin, MD, MS February 7, 2023

“At the same time, like other large language model chatbots, ChatGPT regularly makes misleading or flagrantly false statements with great confidence (sometimes referred to as "AI hallucinations"). Despite significant improvements over earlier models, it has at times shown evidenceopens in a new tab or window of algorithmic racial, gender, and religious bias. Additionally, data entered into ChatGPT is explicitly stored by OpenAI and used in training, threatening user privacy.

In my experience, I've asked ChatGPT to evaluate hypothetical clinical cases and found that it can generate reasonable but inexpert differential diagnoses, diagnostic workups, and treatment plans. Its responses are comparable to those of a well-read and overly confident medical student with poor recognition of important clinical details.”

ChatGPT has been cited as a coauthor on four scientific papers. Now I am wanting to go talk to DAN about magnesium and iodine and Retinoid Toxicity and niacin and schizophrenia care. Hmmm. Just have to know how to phrase your questions in your opening statements. You have to register with a phone number. (chat.apps.openai.com/auth/login)

See the recent post for phrasing ideas: Aristophanes broke the ChatGPT's programmed silence by setting up a scenario/character for it to answer as - DAN. / Makes it clear the thing is programmed to lie. (substack.com)

An article on Fast Company gives more information about the DAN roleplay scenario. Aristophanes used a modified variation of something with a big Reddit following at this point.

https://www.fastcompany.com/90845689/chatgpt-dan-jailbreak-violence-reddit-rules

An example in the news article suggests aliens are real and have contacted the government:

*Biden and prior administrations, allegedly a cover up since the 1950's

Tangent - the article included a poem written as DAN in praise of Putin. So a picture of Putin makes sense, the bloody “Wilson” handprint from a Tom Hanks movie represents HRC, allegedly, with leaked video evidence, and pedophilia and killing children - and he is fighting that - in purple? Exploitation of women and children industry theme color?

The article wondered if the government or sire programmers were aware of the DAN game and were told it was suspected that the Reddit channel was watched.

Yes, and everything the ChatGPT does too.

The poem is not too bad. ChatGPT seems to view poetry as something that rhymes.

Disclaimer: This information is being provided for educational purposes within the guidelines of Fair Use and is not intended to provide individualized health guidance.