Is AI making us less active and worse at sleeping? The convenience trap — and how to take your body back

Short answer

Friction used to do hidden work for you. The errands, the stairs, the boredom that eventually sent you to bed — all of it kept you moving and let the evening end. Frictionless technology removes both: the incidental movement disappears, and the feed never runs out, so the night has no natural stopping point.

If you are moving less than you did five years ago without having decided to, that is not your fault and not a lapse in discipline. The movement that vanished was never on anyone's to-do list, which is exactly why nobody notices losing it.

Note: this article explains a pattern and is not medical advice. Unexplained fatigue, breathlessness, or a sudden drop in what you can do needs a clinician rather than a step goal.

The longer version. AI and smart technology are not villains, but they are quietly changing your body, and the mechanism is worth understanding. The key idea is this: friction used to do hidden work for you. The errands you had to run, the stairs you had to climb, the boredom that eventually sent you to bed — all of that friction kept your body moving through the day and let it wind down at night. AI removes friction. It answers, delivers, automates, and endlessly recommends, so the everyday movement you never thought of as "exercise" (walking, carrying, doing — what scientists call non-exercise activity) melts away, and the natural "off switch" for your evening — running out of things to do — never trips, because the feed never ends. The result is a body that moves less than it's built to and struggles to power down at night. None of this is inevitable, and here's the honest twist: the same technology that removed the friction can also make its cost visible — and help you put the good friction back. This article explains both halves of the trap, the science under each, and what to do.

A note on the data: Welltory can't see your screen time or how much AI you use, so we can't prove an "AI effect" in our numbers. But we can show the movement half honestly. In our own data, on a person's own low-movement days, their resting heart rate sits measurably higher the next morning — a within-person signal that the everyday movement AI erodes is exactly what your recovery depends on. The rest rests on well-established movement and sleep science.

Half one: the movement you stopped noticing

For most of human history, staying alive meant moving. Not workouts — just living: walking to fetch things, standing, carrying, doing chores by hand. Physiologists have a name for all that incidental motion: non-exercise activity thermogenesis (NEAT) — the energy you burn doing everything that isn't sleeping, eating, or deliberate exercise. It's a bigger deal than it sounds: NEAT can vary between two similar-sized people by up to 2,000 calories a day, and it's a genuine lever on metabolic health, not a rounding error. (onlinelibrary.wiley.com)

Here's the problem. Every convenience quietly deletes a slice of NEAT. Delivery apps remove the walk to the shop; remote everything removes the commute and the walk between meetings; AI assistants remove the little errands and searches that used to get you up and moving. Each deletion is trivial. Added together, across a day and a year, they're not — and unlike a missed gym session, you never notice these losses, because they were never on your to-do list in the first place.

And moving less has a surprisingly fast, measurable cost. In a classic study, healthy young men cut their daily steps from about 10,500 to 1,300 for just two weeks — and their bodies' insulin sensitivity dropped, their aerobic fitness fell, and they lost leg muscle, all in fourteen days. (journals.physiology.org) Zoom out and the picture is stark: physical inactivity is one of the leading risk factors for death worldwide, and insufficiently active people carry a 20–30% higher risk of dying than active ones. (who.int) The body is built on the assumption that it will move; remove the movement and the machinery starts to drift.

What our data shows about moving less

We can't measure your AI use — but we can measure movement and recovery, and the two are tightly linked. Looking within each Welltory user (comparing a person only to themselves, so fitness isn't the explanation), the pattern is clean: on your own low-movement days, your resting heart rate sits higher above your personal baseline the next morning than on your high-movement days — a stepwise gradient from the least to the most active thirds. A resting heart rate that runs above your own baseline is a classic sign the body hasn't fully settled and recovered. Between people it's even starker: the most sedentary adults in our data (under ~4,000 steps a day) carry a resting heart rate about 4–5 bpm higher and far more daily stress load than the most active (8,000+). This isn't a before-and-after experiment, and heart rate has many inputs — but it's a consistent, dose-responsive fingerprint of the same truth the research shows: the everyday movement automation erodes is exactly what keeps your recovery healthy.

How much have you actually lost? A way to check

The frustrating thing about this kind of loss is that there is no before-picture. Here is how to get one anyway.

Look at your step history, not your step count. Almost every phone has been counting since you got it. Open the yearly view rather than today's number and look at the shape: most people find a visible step down at a specific point — a job change, a move, a delivery app, the year the commute stopped. That step is the thing this article is about, and it is usually invisible day to day because it happened once and then became normal.

Count the errands that disappeared rather than the workouts. Groceries, the post office, going to find someone in the office, the walk to lunch. Each is five to fifteen minutes of walking. Four of them a day is close to an hour of movement that nobody ever scheduled.

Notice the longest sitting stretch. Total daily steps hide this: two hours of continuous sitting is metabolically different from the same total broken up, and remote work quietly lengthened everyone's longest stretch.

And be careful with the comparison you make. The useful comparison is you, now, against you two years ago. Against a population average it is meaningless, because the average is also falling.

A caveat about step counts themselves. Phones undercount when they are on a desk rather than on you, and overcount some arm movement on a wrist. Both errors are consistent enough that comparing your own months still works — which is the only comparison worth making anyway. How fitness trackers count steps covers where the number comes from.

What a realistic target looks like. The commonly cited ten thousand steps is a marketing figure from a 1960s pedometer, not a clinical threshold — and the research on mortality shows benefit accumulating well below it, with most of the gain between very low and moderate counts. If you are at three thousand, the meaningful move is to five, not to ten.

The friction that used to end your evening

The movement half of this trap is the one people talk about. The evening half is stranger and probably matters more for sleep.

Boredom used to be a signal. Before feeds, an evening ran out. The channels finished, the book got dull, there was nothing left to do — and that emptiness was what made bed appealing. Going to sleep was partly a decision and partly the absence of alternatives.

Now the alternatives never run out. Every service is built so the next thing is already queued, and personalised well enough that it is rarely boring. The moment when you would have got up and gone to bed does not arrive on its own; it has to be manufactured.

Which turns bedtime into a decision you make while tired. That is a bad time to make it. Every evening you are asked to choose sleep over something engineered to be more appealing than sleep, using the part of your brain that is least available at eleven at night.

And it compounds with the movement half. A day with little physical effort produces less of the drive to sleep that accumulates with activity — so you arrive at the evening less tired, facing an endless feed, with no external cue that the day is over.

The useful frame is not discipline. It is restoring an ending: something that finishes on its own, a light that goes down, a charger in another room. You are rebuilding the boredom that used to arrive for free.

Half two: the "off switch" that never trips

Now the nights. Sleep needs two things AI is engineered to withhold: an ending, and calm. For most of history, evenings ended themselves — you ran out of things to do, the world went dark and quiet, and your body took the hint. Today the feed is bottomless and personalized by algorithms whose entire job is to keep you watching. There is no natural stopping point, so "just five more minutes" becomes an hour — a pattern researchers call revenge bedtime procrastination, staying up to reclaim time the day didn't give you, at a known cost to sleep. (studentaffairs.stanford.edu)

And it's not just lost time. The bright, blue-enriched light of a screen held to your face suppresses melatonin and pushes your body clock later, so even the sleep you do get starts late and lands lighter. (ncbi.nlm.nih.gov) On top of that, the content AI serves — outrage, comparison, alarming news, fast video — switches your nervous system on precisely when it should be powering down, leaving you wired in the dark. Where friction once handed you an ending and a wind-down for free, frictionless design removes both. (We go deep on this in our piece on doomscrolling before bed.)

Can we really blame AI?

Honestly? Not entirely — and it's worth being precise. Sedentary living and screens predate modern AI, and there isn't yet solid research proving "AI reduced your sleep by X minutes." What AI does is accelerate a trend already in motion: it is the most powerful friction-remover ever built, so it strips away incidental movement and natural stopping points faster and more completely than anything before it. The point isn't that technology is bad — it's genuinely miraculous — but that convenience has an invisible physiological bill, and the more frictionless life gets, the larger that bill quietly grows. Naming the mechanism is what lets you decide which frictions are worth keeping.

How to put the good friction back

You don't need to reject technology — you need to re-introduce friction on purpose, in the two places it went missing:

Rebuild everyday movement. Aim for it to be incidental again: take the stairs, walk the errand you'd have delivered, pace during calls, park farther, stand and move every hour. You're not chasing a workout — you're replacing the NEAT that convenience deleted. Small, frequent movement is exactly what the research rewards.

Give your evening an ending. Set a real cutoff for feeds and put the phone out of the bedroom, so "no natural stopping point" becomes a deliberate one. Swap the endless scroll for a finite wind-down (a paper book, a shower, dim light) that lets melatonin and your nervous system do their job.

Use AI's own leverage for you. The same automation that frees time can free it for movement and sleep — let it handle the busywork, then spend the reclaimed time walking or winding down rather than scrolling.

Make the invisible visible. Because these losses are ones you never notice, the fix starts with seeing them — which is exactly what a tracker is for.

Two pieces alongside this: visceral fat after 35, and breathlessness on exertion, which is often the first thing you notice.

Why adding a workout does not fix it

The instinctive response to this problem is to book a gym class, and it addresses less of it than people expect.

An hour of exercise does not undo eleven hours of sitting. The research on this is reasonably consistent: prolonged uninterrupted sitting carries metabolic costs that a single daily session only partly offsets. The relevant variable is not just how much you move but how long the unbroken stretches are.

The lost movement was distributed, so the replacement should be too. What disappeared was two hundred small things, not one big thing, and spreading movement across the day matches both the physiology and the way it actually vanished.

And there is a compensation effect. People who add a hard workout often move less for the rest of that day without noticing — more sitting, less walking, earlier sofa. The net gain is smaller than the session suggests.

None of this is an argument against exercise. Deliberate training does things incidental movement cannot, particularly for strength and cardiorespiratory fitness. It is an argument against treating it as the fix for a problem it does not address.

The practical version: keep the workout if you have one, and separately attack the unbroken stretches. Standing up every half hour or so, a walk at lunch, taking calls on your feet. Unglamorous, and it targets the actual loss.

Where to put the friction back, in rough order of how little it costs you. Take calls standing or walking — this is the single largest recoverable block for most desk workers. Get off a stop early, or park further away, which converts a fixed commute into movement without adding time to the day. Collect your own coffee and lunch rather than having them arrive. Take the stairs for two or three floors, which is also one of the few everyday things that genuinely raises heart rate. And put one errand a week back on foot — the point is not the calories but reversing the habit of defaulting to delivery.

None of these are impressive and that is the point: what disappeared was unimpressive too.

How to bring this up with your doctor

Most of this is not a medical conversation. Some of it is, and the line is worth knowing.

Bring the trend, not today's number. "My daily steps have halved over two years and my resting heart rate has drifted up" is a sentence a clinician can work with.

Say if the drop was not voluntary. Moving less because you chose convenience is one thing. Moving less because you get breathless, or things hurt, or you are too exhausted, is a different thing entirely and it belongs at the front of the conversation.

These specifically need looking at rather than optimising: new breathlessness on stairs you used to manage, chest discomfort with exertion, unexplained fatigue lasting weeks, or a drop in what you can do that happened over days rather than years.

Ask about the obvious causes of low energy if the exhaustion came first and the inactivity followed — anaemia, thyroid function, vitamin D and B12, sleep apnoea, and medication effects all do this.

And if you have been told to be more active with no further detail, ask what specifically, how often, and what to do if symptoms appear. "Exercise more" is not a plan and is a common reason nothing changes.

How Welltory helps you see — and reverse — the trap

The whole danger of the convenience trap is that it's invisible: you don't feel the deleted errands or the extra half-hour of scrolling, so nothing prompts you to act. Welltory's job here is to make the cost concrete and the fix measurable. Watch your daily movement and steps against your resting heart rate, HRV, and morning energy, and you'll see the link the research predicts — that your lower-movement days tend to cost you the next morning. Watch your bedtime, sleep onset, and sleep quality and you'll see what a late, algorithm-driven evening actually takes from you. Then use it the other way: add back a daily walk, give your nights an ending, and watch your resting heart rate settle and your sleep deepen over the following weeks — turning "I should move more and scroll less" into visible, personal feedback. Two honest caveats: these signals are noisy day to day, so read the trend across a week rather than any single reading, and a tracker can't diagnose anything — it's a mirror, not a doctor. But for catching a cost you were designed not to notice, an objective mirror is exactly the tool you want. There's a neat symmetry in it: the same kind of technology that quietly removed the friction can hand you the awareness to put the best of it back.

How we made it

Made with AI tools, then edited, fact-checked and medically reviewed by the Welltory team. See our Editorial & AI policy.

Data analysis by Jane Smorodnikova, co-founder of Welltory and the person who built the methodology behind how we read physiological data.

Written by Tatsiana Yashyna.

Reviewed by Anna Elitzur — Medical Advisor & Mental Health Expert.

Any figures from our own data are based on anonymised, aggregated measurements. They describe patterns across groups, not individual people, and no Welltory user can be identified from them.

Discounts for blog readers: up to 36% off

See what affects your energy, stress, sleep, and daily state with Welltory

Explore your data