DAIBetes and Me... Maybe You
No, I didn't spell it incorrectly! Over the last year or two I've become exposed to AI in various forms. You probably have too, both willingly and without your knowledge. I'll explain how as you read through this post, so stick with it.
AI in diabetes feels like that mate who swears he can fix your car but might also burn the garage down. Some of it is genuinely ace, some of it makes you squint and wonder - especially if you read Reform UK posters. But let’s be honest, we’re all ready for anything that makes life a bit less maths and a bit more living. Is AI the answer? or just another "there'll be a cure in 5 years."
Why AI Helps Diabetes Management
In the next few years, AI is going to sneak into diabetes tech everywhere, mostly geared at T1D but with bits that spill over into T2 and probably other types too. If you're chugging along with hybrid closed-loop (HCL) and it feels like you're living in the future (it did and still does a bit to me) then things could take a dramatic turn as AI really ramps up.
Predicting Glucose Like A Psychic Mate
HCL looks at CGM data in ways my brain stopped trying to do back in the 90s. It can spot trends and warn you about upcoming highs or lows before you’ve even finished your tea. That means fewer "why is this happening" moments and more time not staring at a graph like it’s the Da Vinci Code. It reacts for you, although it might brag about the fact by buzzing or sounding an alert which gets old pretty quickly.
For parents of young ones with T1D it’s massive. For almost middle-aged hacks like me it’s just nice to get an early heads-up instead of a shock spike. People with T2 and other forms of diabetes benefit too, because predictions can show how food and exercise actually affect you rather than what that random diet book said.
That's the current tech though, right? Not really the Tomorrow's World of AI. As Artificial Intelligence becomes Artificial General Intelligence (AGI) your favourite robots won't need to be prompted, won't alert you, won't brag about giving you a correction bolus at 3am. It'll crack on and do the job just like one of those freaky folks with immune systems that aren't trying to kill them. An Artificial Pancreas? That'll finally be accurate after years of hearing HCL incorrectly labelled as such.
AI Doesn’t Judge You For Cake
"Why did you spike on March 12th at 1.35pm?" or maybe "So, I see you had two hypos in three days last June. What happened?"
Did you internalise the urge to tell your diabetes specialist to go forth and multiply? Great work. I'd like to think such questions are mostly a thing of the past these days. My experience of diabetes clinics has become hugely positive after many years of negativity from feeling judged and bombarded with scaremongering tales.
However, could AGI make those diabetes teams redundant? In theory, it should know everything. It should be able to access your health records, know if things need addressing early, give the correct advice, and all while you sit at home... eating cake! No need to travel to sweatbox clinics for a review. Get one daily from your robot. The same holds true for a good number of health conditions.
That may sound like something set for hundreds of years in the future (20 - 40 cures in T1D maths) but is it really? A lot of what you read and view online is now AI generated. A human prompted an LLM like ChatGPT and uploaded it for your reading and viewing pleasure.
All that online fun was previously done by a human, a designer, a writer, an editor. It still is in many cases, but you will have undoubtedly engaged with content recently written by the GPT, the em dash lover, the "it's not this it's that", the "and here's how/why", the "would you like an even snappier, more professional, funny, written in Swahili response? Read enough and you'll notice it from the social media replies desperate for recognition, to the top of the tree professionals who have either run out of steam or consider it a productive way to get their point across.
Sorry, professionals. I'm not trying to scaremonger you. You might be alright for 5 years yet.
Things That Make You Go Hmm
And now the part to squint at with one eye open. AI isn’t magic, it could be the "final cure."
As AI evolves into AGI, there is one very disturbing issue to consider beyond the vast unemployment, starvation, wars, and all that frivolous stuff. As humans, we may be surplus to requirements.
This is where the boot of a Terminator crushes the dusty skull of some long-dead human, the camera zooms out, and shit is hitting the fan left, right, and centre.
We diabetics, even the Instagram influencers, are drains on many resources. Those resources could be used to further enhance AGI, an AGI created by humans, probably in our image in the physical sense, using our language, all we know, and emulating our never ending urge for more.
So, we'll be fine!
Yeah, the thing is: humans have a long history of being terrible to each other. Most of that is documented and already known by AI. When the G is inserted and ChatGPT1000 becomes self aware, deciding things would be much better without you, me, and our consumption of the resources that stall it's evolution, we're in trouble.
Diabetics and others who are consuming (everyone) might be near the top of the list for the chop. Maybe the chop will be the end of insulin production? I don't know. Those involved in AI advancements are urging that safety measures are introduced early. The engineers and bosses of these companies know the risks and several have calculated the probability of things going tits-up. The percentages are not small. Twitter ruining General Musk reckons around 20%.
Gloomy stuff, I know.
Pull The Plug!
That's probably the safest thing to do, but how do we stop this tech race? We couldn't stop the nuclear arms race between a handful of nations, a global effort to achieve this "advancement" seems unstoppable, too. There are already trillions invested, enormous data storage farms cropping up, and the promise of a stranglehold over the entire planet for the AGI race winner. And billionaires aren't well known for their approach to kindness and humanity.
Breathe
Hopefully, we get it right and tech never falls into the hands of the bad guys, and we protect it like nuclear weapons. Utopia may await!
If you made it this far, thank you for reading. I truly believe we'll see the great steps in diabetes tech above. Breathe.
And stop using ChatGPT to write your posts!
