image_printPrint Page

AI versus Human: Assessing AI’s Effectiveness in Diabetes Education

Are computers smarter than humans?

As we look at where AI is going, there has been a visible ramp-up in use particularly in medicine. Looking at just one area of health and medical use, diabetes stands ripe for application. But how far are we and where do we need to get before we can really trust the evaluation and recommendations from a non-human system?

I recently read an article called “ChatGPT passes the nutrition test, but experts remain irreplaceable” by Tarun Sai Lomte that piqued my interest in looking further at how well AI is “learning” for education purposes and what pros and cons need to be considered.


This article goes into some detail about when AI is given the task of providing education.

Where AI does a great job is in navigating one type of education need.

Give it direction to provide a low-carb diet, a vegan diet, or a low Cholesterol diet, and it seems to do a good job of providing that information and even personalizing it just a bit. At this point it could be told to provide a diet plan for a certain number of calories or with exercise in mind as well as for low sodium–it can direct it that way.

However, where diabetes is concerned within the realm of nutrition, there are many health conditions a person might be managing along with medications and personal preferences.

There might be digestive issues, kidney, or heart complications. Mixed into this could be food allergies or intolerances, nutrient deficiencies, or just simple food aversions. When I consider where AI is at this point, it seems to falter a bit in terms of directing the information for so many things at one time. The system can’t get as specific as needed because unfortunately some of the information for one condition, type of medication or preference may be counter to what the other condition should need.

Just as diabetes medications have become more complex, so have the medications and recommendations for other health conditions. When a system is given information to use for one specific issue it may have a hard time applying the human side of managing it all. Directions may be very machine-like in the delivery method and that leaves many people feeling like they could have just gone to a search engine and gotten the same information.

On the other side of considerations are some of the pros that can come from using AI in healthcare settings.

What if it could help doctors or other clinicians make decisions in less time? Think of all the time that could be saved in a visit with your doctor IF AI had already looked at lab work, medications, and glucose/pump records. With all the guidelines that must be followed, it would be nice for some of that to be put together to allow the doctor more time to discuss what your needs are as well as what needs to be adjusted.

Consider it from another angle. Looking at all the other health checkups that are necessary with diabetes, electronic systems could pop up reminders when someone needs a visit with an ophthalmologist or a podiatrist based on info in the chart. The doctor could ask about these things but may not with time constraints. If the system checked all these things automatically the doctor could spend more time in discussion about the things that matter and direct a person to make appointments for those that are overdue.

With so many things to consider in terms of using AI, if the clinician has something that’s bringing enough data together to provide direction for that visit and some ability to personalize the visit at a human level it will make a difference in motivating the person with diabetes.

All in all, we have a lot to be thankful for with progress to using AI with diabetes.

Eventually, it may be able to tell us when we last ate at the local pizza shop, what we did and how to adjust for it to ensure a better outcome this time. It isn’t close to being perfect and we still need a human angle when it comes to personalization and the intricacies of life with diabetes. If you need the human angle of education, as always, the clinicians at Integrated Diabetes can help!

image_printPrint Page