AI learns from examples
AI systems are trained using large amounts of data. During training, the system looks for patterns.
For example:
- An AI system trained on text may learn how words and phrases commonly appear together.
- An AI system trained on images may learn visual patterns.
- A medical AI tool may be designed to help analyze certain healthcare data.
This does not mean AI understands your health the way your doctor does. It means the AI has learned patterns from data.
AI uses your prompt and context
When you type into an AI chatbot, the system breaks your words into smaller pieces called tokens. A token may be a word, part of a word, or even a single character.
The AI then uses those tokens to generate a response. Many language AI systems estimate what token or sequence of tokens is likely to come next, based on the prompt and context.
AI breaks language into small pieces
Your prompt is what you ask the AI.
Your context is the information the AI can see at that moment, such as:
- Your question
- Instructions you provide
- Text you paste
- Files or information you upload, depending on the tool
Better context can lead to more useful answers.
Instead of asking:
What should I ask my doctor?
You could ask:
I have CLL and I am preparing for a visit. Please help me make a list of questions. Focus on questions I can discuss with my healthcare team.
AI generates an answer, but it does not guarantee accuracy
AI-generated answers can sound clear and confident even when they are incomplete or wrong.
For health questions, this matters because AI may not know your:
- Full medical history
- Current lab results
- Medications
- Other health conditions
- Treatment goals
- Personal preferences
Better context can lead to more useful answers.
Mayo Clinic advises that AI should not replace healthcare professionals and should not be treated as a secure place to enter sensitive medical or personal information.
AI can help you prepare for shared decision-making with your healthcare provider
AI may be useful as a preparation tool. It can help you:
- Organize questions
- Simplify medical terms
- Summarize topics you want to discuss
- Think about your goals and preferences
Shared decision-making means patients and healthcare professionals work together to discuss options, benefits, risks, and what matters most to the patient.
Should I have, or confirm that I have already had, testing for important CLL markers such as del(17p) by FISH testing, TP53 mutation, IGHV mutation status, and any other tests my healthcare team recommends?
How would del(17p), TP53 mutation, unmutated IGHV, or other results affect my prognosis, monitoring plan, and treatment options?
Which treatment options have published long-term follow-up of 6 years or more, and what do those data show about how long responses may last, the chance of CLL returning or progressing, overall survival, and long-term side effects?
Which treatment options have evidence in people with del(17p), TP53 mutation, and/or unmutated IGHV?
What are the possible benefits, side effects, monitoring needs, infection risks, heart risks, or other important safety considerations for each option?
Is the treatment continuous or fixed-duration? How often would I need visits, hospitalizations, lab tests, scans, or monitoring?
How might each option fit with my goals, such as staying active, reducing side effects, limiting clinic visits, avoiding hospitalizations, maintaining work or family responsibilities, and preserving quality of life?
What symptoms, lab changes, or other signs would mean that treatment should start or that my current plan should change?
Please organize the output into five sections: Testing Questions, Long-Term Outcomes Questions, Higher-Risk CLL Questions, Safety and Lifestyle Questions, and Shared Decision-Making Questions.
