Necessary
Required for secure sessions, sign-in, CSRF protection, language selection and saving your cookie choice.
← Back to blog
What can an AI assistant help with, and where should it stop?
Keep information and preparation support separate from an individual professional assessment.
An AI assistant may explain how a service works, organise questions or suggest a general professional category to explore. Fluent conversation does not mean it fully understands you or is necessarily correct. It can present inaccurate information with confidence.
Keep the task small and clear
“List five questions for my first appointment” is easier to check than “Fix my life.” Explain the purpose and limits of the result. Asking for a draft is not the same as giving the system authority to decide.
Set boundaries around health decisions
A chatbot should not replace diagnosis, medication decisions or an urgent risk assessment. A response describing symptoms is not a personal examination. For immediate health or safety concerns, seek suitable local help rather than waiting for the assistant.
Make answers verifiable
Ask it to separate established information from uncertainty. Open the sources yourself; names and links can be fabricated. Verify platform prices, availability and cancellation conditions on the actual relevant pages. Unknown information should not be guessed and presented as fact.
Move to human support when needed
If answers repeat, the question is misunderstood or individual assessment is required, speak to a person. This is an appropriate boundary, not something you must overcome. AI may be a useful preparation tool, but you do not have to leave responsibility for decisions or care with a chat window.
Sources and further reading
NIST — Generative Artificial Intelligence Profile (AI 600-1)