Skip to content

Insights

Can I Put Patient Information Into ChatGPT? A Simple Guide for Healthcare Staff

The issue is not only what an AI tool produces. It is also what information is entered into it.

06PRACTICALFAIRE

General educational information and FAIRE’s interpretation of the linked sources. This is not legal, clinical or regulatory advice. Read the disclaimer.

Pause before entering information

Patient names, contact details, medical record information, clinical notes, photographs, screenshots, referral documents and identifiable clinical histories should all make staff stop and think before using an AI tool.

Removing a name does not automatically guarantee that information can no longer relate to an identifiable individual. Context and combinations of details can matter too.

A simple principle

If you would not put the information into an unapproved external service, do not put it into an AI tool simply because the interface looks familiar. The same care should apply whether the task is drafting, summarising or asking for a suggestion.

Practices should establish which tools and uses are permitted rather than leaving every staff member to make their own decision in isolation.

What should staff ask?

  • 01Is this tool approved for this type of information?
  • 02Could the details identify a patient directly or indirectly?
  • 03Is there a safer way to complete the task without sharing the information?
  • 04Who can I ask when I am unsure?

FAIRE takeaway

Familiarity is not approval. Clear, practical guidance helps staff pause at the right moment and protects patients, professionals and the practice.

Need more practical guidance?

FAIRE helps healthcare practices understand, govern and implement responsible AI.

Request support →