Learning pathology: Difference between revisions

Jump to navigation Jump to search
Mikael Häggström (talk | contribs)
Mikael Häggström (talk | contribs)
 
Line 129: Line 129:


==Using AI "chatbots"==
==Using AI "chatbots"==
For AI "chatbots" (such as ChatGPT, Gemini or Copilot), you can generally trust a simple answer to a simple and presumably common question. More complex answers generally require you to provide more details together with your question. For example, a question of "what is the risk that this patient has..." would require you to provide a patient summary for a better answer, while avoiding identifiable information.<ref group=notes">See [https://www.hhs.gov/hipaa/for-professionals/special-topics/de-identification/index.html#standard HHS de-identification standards] for what counts as identifiable information</ref>
For AI "chatbots" (such as ChatGPT, Gemini or Copilot), you can generally trust a simple answer to a simple and presumably common question. More complex answers generally require you to provide more '''details''' together with your question. For example, a question of "what is the risk that this patient has..." would require you to provide a patient summary for a better answer, while avoiding identifiable information.<ref group="notes">See [https://www.hhs.gov/hipaa/for-professionals/special-topics/de-identification/index.html#standard HHS de-identification standards] for what counts as identifiable information</ref> '''Avoid leading''' prompts (such as "these features indicate invasiveness, right?") but rather keep them neutral and open-minded.


==Using question banks==
==Using question banks==