Are Advancements in Diabetes Technology Truly Helping Us?
By Rabbi Hirsch Meisels, Director, Jewish Friends With Diabetes International
- How can we maximize the benefits we can get from them?
- How can we maximize the benefits we can get from them?
Walk into the world of diabetes today and you will see a landscape that would have seemed unimaginable just a few decades ago. Continuous glucose monitors display data every few minutes. Insulin pumps adjust delivery automatically. Advanced closed-loop systems, also known as automated insulin delivery (AID) systems, promise smoother glucose curves and fewer dangerous extremes. For anyone who remembers the earlier era of diabetes care, the contrast is staggering. For many, diabetes management now looks more like software than medicine.
I still remember growing up with diabetes in an era before any form of glucose monitoring existed. You did not “see” your blood sugar. You guessed, and you learned to read your body’s signals. Anyone diagnosed today, entering a world of sensors, smart pumps, and algorithms, can hardly imagine what that reality was like. In this sense, technology has unquestionably transformed diabetes care. It has given us tools that can prevent many severe lows, shorten the duration of highs, and reduce a tremendous amount of daily burden.
And yet, alongside this progress, an important question quietly emerges:
Are these technological advancements automatically working in our favor — or are we at risk of losing something essential along the way?
Is Technology Powerful — But Autonomous?
Modern AID systems adjust insulin delivery based on continuous glucose readings. This is an extraordinary achievement. However, these systems do not function in isolation. Their performance depends heavily on the quality of information they receive from the user — especially around meals.
A recent study published in Diabetes Care compared simplified meal announcements with precise carbohydrate counting in adolescents using an advanced hybrid closed-loop system. The conclusion was clear: accurate carbohydrate counting led to significantly better glucose outcomes, even with one of the most sophisticated commercial algorithms available.
In other words, even the smartest system still performs better when the human side does its part — a reality many clinicians and experienced users already see in everyday practice.
A diabetes educator once summarized this simply: a pump is only as good as its user.
Two Emerging Approaches Among Users
Today, we can roughly observe two different styles of using AID systems.
1. The “mostly hands-off” approach
These users place strong trust in automation. They may glance at glucose trends and roughly estimate meals, but they largely expect the system to “figure it out.” When glucose rises, they assume the algorithm will correct it. When a meal is misjudged, they wait.
The practical result?
Often, the pump does eventually bring glucose back toward range — but sometimes only after hours of elevated levels.
I have witnessed a similar pattern before. Years ago, people transitioned from rigid fixed insulin regimens to flexible basal-bolus therapy with carbohydrate counting — a shift that promised greater freedom and better control.
In reality, while freedom increased, overall control often worsened rather than improved. The tools became more powerful, but they also demanded more responsibility. Not everyone was prepared for that shift.
We may now be watching a similar phenomenon as people move from simpler pumps to more automated systems. The technology is more advanced, yet in some cases overall control deteriorates — not because the devices are worse, but because increased automation reduces active engagement.
2. The “technology-as-a-tool” approach
These users view automation as an assistant, not a replacement. They:
- Count carbohydrates carefully
- Watch trends
- Fine-tune boluses
- Intervene when needed
They benefit from automation while still exercising human judgment.
In real-world experience, this second approach consistently produces tighter control. The difference is not the technology — it is the level of human engagement, and the willingness to remain an active participant rather than a passive observer.
Are We Quietly Losing Essential Skills?
A growing trend in device development is pushing toward systems that minimize or even remove manual input. Some newer designs emphasize meal estimation, simplified announcements, or minimal interaction.
Convenience is appealing. But it raises a serious concern:
If a person never needs to calculate carbohydrates or think through insulin dosing, how will they ever learn how diabetes truly works?
Even for those who already possess these skills, relying on systems that discourage manual involvement can gradually weaken them. Human nature gravitates toward the path of least resistance. If effort is no longer required, effort eventually disappears.
A friend of mine, who had been very actively controlling his diabetes for years, recently switched to a system with minimal manual capability. When I asked him why, he answered, “I needed a vacation from my diabetes.”
A vacation is understandable. But a four-year vacation?
We often hear the claim that smartphones, while incredibly useful, have made many people less mentally engaged — because so much thinking is now outsourced to devices. A similar dynamic may be developing with smart insulin pumps.
This is not a minor issue. Manual understanding of diabetes physiology, insulin action, and food impact remains the foundation of safe self-care.
The Education Gap Created by Technology
This problem extends beyond skill erosion. It also affects how new patients are trained from the very beginning.
Beyond losing skills lies a deeper problem: the loss of education pathways.
Many newly diagnosed patients now start directly on CGM and automated insulin delivery pumps. They never go through a phase of learning basal/bolus logic, correction factors, carb ratios, or pattern management. This is not a critique of CGM or AID technology itself, but of systems and care models that eliminate meaningful manual user input. When something breaks, fails, or behaves unexpectedly, these users often lack the conceptual understanding of what to do.
This means the problem is not only that people become lazy — it is that systems unintentionally create users who were never taught the fundamentals.
Meaningful diabetes control requires time, education, and sustained investment in understanding fundamentals such as carbohydrate counting, insulin action, and pattern recognition. No technology can replace that foundation.
Technology Can Mask Problems Instead of Fixing Them
Another subtle but powerful idea:
Automation often covers errors rather than correcting their cause.
Examples:
- Poor carb estimation
- Inaccurate ratios
- Inappropriate basal settings
- Suboptimal diet patterns
For example, a person may routinely underestimate carbohydrate intake at dinner without realizing it. The pump partially compensates, the glucose curve eventually flattens, and no alarm is triggered. The user concludes, “Everything is fine,” while in reality the same error repeats night after night — quietly limiting the quality of control.
The pump “smooths” the curve, but underlying mistakes remain. Over time, users may believe control is good, while variability and hidden instability persist.
This reinforces:
Technology can hide weaknesses.
Human involvement discovers and fixes them.
Psychological Ownership
When people feel they control diabetes, they develop responsibility and awareness.
When a device feels like it controls diabetes, people emotionally disengage.
This explains behavior beyond laziness — it represents a shift in identity:
“I manage diabetes” → “The pump manages diabetes.”
That transition can be dangerous in the long term.
Can Safety Algorithms Become a Barrier to Excellence?
An important limitation of many modern pump systems is how conservative their built-in safety logic has become. These devices are designed to strongly favor avoiding hypoglycemia, and when faced with a choice between a potential low or a higher glucose level, they will almost always choose the higher number — which is, of course, critically important. But in practice, this often means that the system enforces its own definition of “safe” glucose targets, regardless of an individual user’s goals or circumstances.
Some pumps will automatically reduce or suspend insulin delivery if they predict glucose will drop below a certain threshold — even if the user intentionally aims for tighter control, and even if the person knows from experience that their body can safely tolerate lower values in specific situations and at specific times.
The result is that perfect or near-perfect control becomes structurally unattainable, not because the user lacks skill, but because the system simply will not allow it.
This creates a frustrating paradox:
The more aggressively a knowledgeable user tries to fine-tune control, the more the pump may “fight back” by cutting insulin.
In real life, slightly lower glucose targets can be reasonable and beneficial for individuals who carefully monitor trends and respond early. Yet the pump’s conservative logic treats all users and all situations similarly.
This highlights an important truth:
Current AID systems are optimized for population-level safety, not individualized excellence.
None of this means these systems are bad. It means they are incomplete. And it reinforces the central message of this editorial: technology is a powerful assistant — but it cannot yet replace individualized human judgment.
A Simple Analogy
Consider the current state of self-driving cars.
A teacher once related to me a story in the teachers’ room. He lives in New York, where driverless cars are essentially nonexistent. One weekend, while visiting Los Angeles, someone ordered him a taxi. He entered what he assumed was a regular vehicle, expecting to see a driver behind the wheel. Instead, he found an empty driver’s seat.
The doors closed, the steering wheel began turning on its own, and the car slowly started moving. For a moment, he froze. Then he panicked. He had no idea how the vehicle could be controlled or stopped, or what was safe to touch. After fumbling around, he finally located a support button and contacted an operator. The operator calmly instructed him to wait. Within moments, the car pulled over to the side of the road and came to a stop. He opened the door and got out — shaken, but unharmed.
One day, driverless cars may be flawless. One day, no one will think twice about stepping into a vehicle without a driver.
But today?
Most people are still uneasy — because we know the technology is not perfect yet.
Insulin pumps are in a similar stage.
They are remarkable.
They are helpful.
They are not yet flawless.
Handing over full responsibility to a system that we know has limitations is not wisdom. It is premature trust.
Caution is not anti-technology. It is pro-safety.
When Will the Story Change?
If, one day, systems exist that can truly maintain near-perfect glucose control across meals, stress, illness, and daily variability — without meaningful human input — then the conversation will change.
At that point, deep manual fluency may no longer be necessary.
We are not there yet.
The Proper Role of Technology
Technology does many things extremely well:
- Reduces frequency and severity of lows
- Dampens prolonged highs
- Handles background insulin delivery more smoothly than humans ever could
These are enormous achievements.
Ultimately, our goal is not simply to avoid extremes, but to achieve glucose profiles that are as flat and stable as possible, in order to reduce long-term complications.
But technology today does not provide consistently perfect control. Sensors have inaccuracies, algorithms make assumptions, and real life introduces constant variables — absorption, stress, illness, hormones, and activity.
To reach truly excellent control, human input is still required.
The Balanced Path Forward
This philosophy should guide not only people living with diabetes, but also educators and device manufacturers.
The healthiest approach today is a middle ground:
- Use the best technology available
- Choose systems that allow meaningful manual input
- Stay actively involved
- Continue learning and practicing core diabetes skills
Technology should support our management — not replace our understanding.
Bottom Line
Advancements in diabetes technology are absolutely for our benefit — if we use them correctly.
They become less beneficial when they:
- Encourage disengagement
- Remove user control
- Replace education with blind trust
The goal is partnership, not surrender.
Machines are improving rapidly.
Humans must not stop improving alongside them.
When we combine advanced tools with educated, engaged users, we get the best of both worlds — and the best possible outcomes.
Watch Gary Scheiner and Rabbi Hirsch Meisels in episode 29 of Think Like a Pancreas–The Podcast, when they talk about Powering Through Passover with Type 1 Diabetes.
Listen as they delve into the challenges of extended mealtimes during the Seder and discuss effective glucose management strategies. Learn about the importance of timing insulin dosages, especially in relation to the Seder’s unique structure, and gain insights into carb counting techniques for matzah and other Passover foods to ensure a balanced and healthy observance.
One Response
Every person living with diabetes on a CGM/CSII/AID system or considering one should read this post, as well as everyone who lives with someone with DM or works in healthcare. Outstanding post.