Tips to navigate AI-generated health information while managing diabetes
I recently had the privilege of presenting at the ADCES (Association of Diabetes Care & Education Specialists) annual national conference on “Understanding and Adapting to Generative AI in Clinical Practice.” As I spoke to hundreds of diabetes educators about the intersection of artificial intelligence and diabetes care, I realized there’s an important conversation we need to have with people managing diabetes and prediabetes.
While AI tools, like ChatGPT, can be incredibly helpful for learning and organizing information, they also present unique challenges when it comes to health decisions. Today, I want to share insights from both sides of this conversation – as someone who teaches healthcare professionals about AI and as a diabetes specialist who wants to keep you safe and informed.
The Promise and the Pitfall 📊
Recent research shows that AI chatbots can outperform physicians in certain areas – they’re often rated as more empathetic and provide more comprehensive responses to general health questions. But here’s what the research doesn’t capture: the nuances that matter most when you’re managing a complex condition like diabetes.
AI excels at processing vast amounts of information quickly, but it struggles with what researchers call “uniqueness neglect” – the inability to account for your individual circumstances, cultural background, medication interactions, and personal health history.
Red Flags: When AI Advice Could Be Dangerous ⚠️
Based on my clinical experience and research into AI limitations, here are warning signs that AI-generated diabetes advice needs professional verification:
Specific Medical and Health Recommendations
- Exact medication dosing suggestions
- Advice to start, stop, or change medications
- Specific insulin-to-carb ratios without knowing your current regimen
- Personalized blood glucose and A1C goals based on health conditions
- Claims about “reversing” or “curing” diabetes
One-Size-Fits-All Solutions
- Generic meal plans without considering your cultural food preferences
- Exercise recommendations that don’t account for your fitness level, health complications, or schedule
- Blood sugar targets that differ from what your healthcare team has set
- Inherent bias from training data that may not represent diverse populations
Overly Confident Claims
- Guarantees about specific outcomes
- Medical advice presented without appropriate disclaimers
- Information that contradicts established diabetes management guidelines
- Inability to critically analyze research data using evidence-based methods
- Confirmation bias – telling you what you want to hear rather than medically accurate information
When AI Can Be Your Learning Partner 💡
AI isn’t all risk – when used thoughtfully, it can be a valuable educational tool:
Safe Uses Include:
- Understanding general diabetes concepts and terminology
- Getting ideas for recipe modifications (that you verify with your dietitian)
- Help organizing questions before your next appointment
- Learning about different types of diabetes technology
💡 Smart Approach:
Think of AI as a research assistant, not a medical advisor. Use it to gather information, then discuss findings with your healthcare team.
Smart Prompting: Getting Better AI Responses
If you choose to use AI for diabetes-related questions, you’ll get much better results using proper prompting techniques. Both OpenAI’s President Greg Brockman’s “anatomy of a perfect prompt” and MIT’s guide to effective prompts emphasize the importance of clear, structured communication with AI.
The GRWC Method
This framework helps you create more specific, useful prompts:
Goal: Start by being crystal clear about what you want to learn. Instead of asking “What’s good for diabetes?” try “I want to understand how traditional Korean side dishes might affect my blood sugar levels.”
Return Format: Tell the AI how you want the information presented. You might ask for “a simple list with explanations” or “step-by-step instructions” or “questions I should ask my healthcare team.”
Warnings: Include important limitations or safety considerations. Mention your medications, health conditions, or specific things you need to avoid. This helps the AI provide more relevant and safer responses.
Context Dump: Share relevant background information. Your cultural food preferences, newly diagnosed status, family situation, or specific challenges all help the AI give you more tailored educational information.
Making Your Prompts Even Better
As you craft your prompts, consider these practical points. Think like you’re explaining to someone who doesn’t know your background, because the AI doesn’t. Mentioning that you’re newly diagnosed or unfamiliar with diabetes terminology will get you more complete explanations.
Be specific about your cultural context rather than just saying “Asian foods.” Mention specific cuisines, ingredients, or cooking methods you’re curious about. Set practical boundaries by mentioning if you only have access to certain ingredients, equipment, or time constraints to keep suggestions realistic and doable.
Most importantly, always build in safety checks by asking questions like “What warning signs should I watch for?” or “When should I contact my healthcare team about this?” Request verification questions by ending your prompts with what you should discuss with your doctor or dietitian. This creates a natural bridge between AI-generated information and professional guidance.
🎯 Better Prompting in Action
Instead of asking: “What should I eat with diabetes?”
Try this approach:Goal: I want to understand how different Asian vegetables affect blood sugar levels.
Return Format: Provide a simple list with explanations of how each vegetable impacts glucose.
Warnings: I take metformin and have high blood pressure, and avoid recommendations for major dietary changes.
Context Dump: I’m Korean American, newly diagnosed with type 2 diabetes, and want to continue enjoying traditional foods while managing my condition. What questions should I ask my dietitian about incorporating these foods?
Notice how the improved prompt gives you educational information while building in the step to verify everything with your healthcare team.
Your Healthcare Team: The Irreplaceable Human Element
What AI cannot provide – and what remains essential for your diabetes management – goes far beyond just having the right information.
Knowledge Isn’t Everything
If knowledge alone could solve diabetes management, we would have solved this challenge long ago. The reality is that managing diabetes successfully requires much more than information. It requires accountability – someone who checks in with you, celebrates your progress, and helps you navigate setbacks. It demands personalization that goes beyond generic recommendations to create strategies that fit your specific life, culture, and circumstances.
Most importantly, diabetes management requires someone with the clinical judgment and critical thinking skills to evaluate whether information is accurate and whether it actually applies to your unique situation. AI can provide general information, but it cannot assess whether advice makes sense for someone taking your specific medications, dealing with your particular complications, or navigating your family dynamics.
What Human Expertise Provides
Clinical Judgment: Your healthcare team can interpret your unique pattern of blood sugar readings, adjust recommendations based on how your body responds to treatments, and recognize when symptoms suggest complications that require attention.
Cultural Competency: Understanding how your family dynamics affect food choices, adapting recommendations to fit your cultural celebrations, and addressing the emotional and social aspects of diabetes management in ways that respect your background and values.
Personalized Care: Coordinating between your various healthcare providers, adjusting treatment plans based on life changes, and providing the ongoing accountability and support that turns good intentions into lasting habits.
Critical Evaluation: Your diabetes care team can assess whether new information you’ve found (whether from AI or other sources) is accurate, relevant to your situation, and safe to implement given your current health status.
Authentic Human Connection: The genuine empathy and emotional support that comes from being truly heard and understood by another person – someone who can recognize the frustration in your voice, celebrate your small victories, and provide comfort during difficult moments in ways that feel real and meaningful.
Guidelines for Smart AI Use 📋
Before Using AI for Diabetes Questions:
- Remember your goal – education and understanding, not medical decision-making
- Check your sources – AI responses aren’t peer-reviewed or verified
- Consider your complexity – the more medications or complications you have, the more you need human expertise
- Prepare questions – use AI-generated information to have better conversations with your healthcare team
After Getting AI Responses:
- Fact-check with reliable sources like the American Diabetes Association, Academy of Nutrition and Dietetics, and the Association of Diabetes Care and Education Specialists.
- Discuss findings with your healthcare team before making changes
- Test carefully if trying new approaches (with professional guidance)
- Stay connected with your diabetes care team for ongoing support
The Bottom Line: AI as Tool, Not Replacement 🔧
AI represents an exciting advancement in how we access health information, but managing diabetes successfully still requires the human touch. Your healthcare team brings clinical judgment, cultural understanding, and personalized expertise that no algorithm can replicate.
Use AI as a learning companion – to understand concepts, explore questions, and prepare for appointments. But when it comes to making decisions about your health, trust the professionals who know you, your history, and your unique circumstances.
Your Next Steps 📅
If you’re interested in learning more about evidence-based diabetes management that honors your cultural background and personal preferences:
Want personalized guidance? Schedule your first session here if you’re using insurance or book a free strategy call to explore how we can work together.
Ready to start making changes? Join my FREE 8-Day Diabetes Transformation Challenge for daily practical tips that respect your cultural traditions while supporting your health goals. Sign up here.
👋 About Karen I’m Karen Lau, a Registered Dietitian and Certified Diabetes Care and Education Specialist with over 15 years of clinical experience helping Asian Americans manage diabetes while honoring their cultural traditions. My background in diabetes care at a major medical institution has given me deep insight into creating culturally-informed approaches to blood sugar management.
Note: This content is for general knowledge and educational purposes only, not medical advice. While this guide follows current medical guidelines, always consult your healthcare team for personalized recommendations and advice specific to your situation. This article was authored by me, with my professional expertise, clinical experience, and original ideas forming the foundation of the content. Claude, an AI assistant from Anthropic, helped organize and expand my thoughts into this comprehensive blog post while maintaining my voice and professional perspective.