AI tools are becoming part of everyday work for many health professionals.
They are used to help summarise information, draft explanations, organise notes and clarify complex topics. When used appropriately, these tools can save time and support clearer communication.
However, as AI use becomes more common in healthcare settings, one basic principle remains essential:
Never enter identifiable patient information into a public AI tool.
This rule is simple, but it is easy to overlook in practice.
Why this matters
Many AI tools available today operate through publicly hosted systems. When text is entered into these tools, the information may be processed on external servers. Depending on the platform and settings, that information may be stored temporarily or used to improve the system.
For this reason, entering identifiable patient information into public AI tools can raise serious confidentiality and privacy concerns.
Even when the intention is harmless, the risk remains.
For example, a clinician might paste a section of clinical notes into an AI tool to ask for a summary or help refining the wording of a document. If those notes include details such as a patient’s name or identifying circumstances, confidentiality may be compromised.
Responsible AI use in healthcare therefore begins with protecting patient privacy at every step.
Common identifying details to avoid
Before entering information into an AI tool, it is important to check whether the text includes any details that could identify a patient.
These may include:
-
Names
-
Dates of birth
-
Addresses or locations
-
Hospital or medical record numbers
-
Contact information
-
Unique combinations of clinical details that could identify an individual
Even partial information can sometimes be enough to identify a patient in certain contexts.
A safer habit: de-identify first
A practical approach is to develop the habit of de-identifying information before using AI tools.
For example, instead of entering:
“Summarise this case: John Smith, a 54-year-old patient from Brisbane with…”
You could write:
“Summarise this case: A 54-year-old patient with…”
Removing names and other identifiable details allows the AI tool to assist with the task while maintaining patient confidentiality.
This small step can significantly reduce the risk associated with using AI in clinical or healthcare communication workflows.
Responsible AI use is often about small safeguards
Discussions about artificial intelligence in healthcare often focus on large ethical questions or technological developments.
In practice, responsible AI use frequently comes down to small, practical habits.
Checking for identifiable information before entering text into an AI system is one of the simplest and most effective safeguards health professionals can adopt.
As AI tools become more integrated into healthcare environments, developing these habits will be an important part of maintaining professional standards and protecting patient trust.
Learn more about responsible AI use in healthcare communication
Understanding how to use AI safely is an important part of modern healthcare practice.
The Responsible Use of AI in Healthcare Communication CPD course provides practical guidance on how health professionals can use AI tools responsibly in communication tasks.
The course covers common risks, verification practices and confidentiality safeguards in real-world scenarios.
It takes about one hour to complete, is free to access, and a CPD certificate is available for $9.99 for those who wish to document their learning.
