Back to the blog Education

We Didn’t Get the Time Back. Now AI Is Making the Same Promise.

Smartphones were supposed to handle the practical stuff so we could spend more time with the people we love. What the research says actually happened, and what it tells us about AI.

By James Martucci, CEO of SideKix Published 30 September 2026 Reading time 10 minutes Category Education

I was at Apple when Steve Jobs was there and the iPhone was new. There was a message he kept coming back to. I’m paraphrasing from memory, but it went something like this: technology exists to get you through what you have to do, so you have more time for the people you love.

I believed it. I was in the building when that promise was being made.

So here’s the question I can’t shake. Did we get that time back?

I went looking for the answer. It isn’t close.

Did we end up connecting with the people we care about?

That was the whole promise, so it’s the first thing worth checking.

The US Surgeon General’s 2023 advisory on loneliness pulled together two decades of federal time use data. Between 2003 and 2020, the average American’s time with friends fell from about 60 minutes a day to about 20. Time spent alone rose from 285 minutes a day to 333. For people aged 15 to 24, in-person time with friends dropped from about 150 minutes a day to 40. That’s a decline of nearly 70%.

It didn’t bounce back after the pandemic. The latest American Time Use Survey, released this June, puts socializing and communicating at 35 minutes a day, down from 41 in 2015. On an average day, only 30% of Americans spend any time socializing at all, down from 38%.

The phone didn’t do all of that by itself. TV, longer commutes, remote work, and a pandemic all played a part. But the device that was supposed to bring us closer arrived right in the middle of that drift, and it kept pulling while we did.

Even when we’re together, we’re not always there. In a 2024 Pew survey, 46% of teens said their parent is at least sometimes distracted by their phone when they’re trying to talk to them. The kids noticed.

We can’t put it down, and we know it.

In 2015, 21% of American adults said they were online “almost constantly.” By 2025 it was 41%. For adults under 30, it’s 63%.

And most of us don’t like it. This month, Pew reported that 53% of Americans say they spend too much time on their phone. 45% tried to cut back in the past year. Only a quarter of them say it went very well. 41% say their phone hurts their sleep.

When half the country wants to do less of something and can’t, that isn’t a willpower problem. It’s a design result.

The numbers for kids make that obvious. A Common Sense Media study tracked teens’ actual phone activity. The typical teen got 237 notifications a day and picked up the phone 51 times. About 5% of those notifications arrived between midnight and 5am.

Economists have tried to measure how much of this we’d choose if we could. A randomized study in the American Economic Review estimated that self-control problems account for 31% of social media use. That’s nearly a third of the time spent scrolling that people themselves wouldn’t have chosen.

Aza Raskin, who invented infinite scroll, has spent years saying he regrets it. His summary is the most honest line I’ve found on this whole subject: “In technology, incentives eat intentions.”

The intentions were good. The incentives won.

What about the anxiety?

This is where people get loud, so I’ll be careful.

The trend lines are real. The share of American adults aged 18 to 29 who say they have or are being treated for depression more than doubled, from 13% in 2017 to 28% in 2026. When Facebook rolled out across US colleges one campus at a time, economists were able to compare students who had it with students who didn’t yet. They found a 20% rise in generalized anxiety disorder, driven largely by comparing ourselves to everyone else.

Internal research leaked from inside the company said the quiet part out loud. One 2019 slide about Instagram read: “We make body image issues worse for one in three teen girls.”

But serious researchers push back, and they deserve to be heard. The National Academies concluded in 2023 that the evidence doesn’t support the claim that social media causes changes in adolescent health across the whole population. Psychologist Candice Odgers argues the arrow may point both ways: kids who are already struggling may use social media more. Poverty, violence, and the opioid crisis matter too. And the latest CDC data shows teen sadness is actually improving, down from 40% in 2023 to 33% in 2025.

Here’s where I land. When researchers pay people to step away, the results are consistent and modest. In the largest Facebook deactivation experiment, people who quit for four weeks were a little happier and less anxious. They also got back about an hour a day.

Where did that hour go? Some of it went to friends and family. Some of it went to watching TV alone.

That’s the whole story in one finding. Freed time doesn’t automatically go to the people we love. It goes wherever the strongest pull is.

Who was paying for all of it?

In 2025, 97.6% of Meta’s revenue came from advertising. About 81% of Alphabet’s did too. Global ad spending passed $1.1 trillion.

When a business earns almost all of its money by selling attention, your time is the inventory. Every minute technology saved you was a minute someone else could sell.

That’s not a conspiracy theory. It’s in the court filings. According to unsealed internal documents reported by TIME, “teen time spent” was set as a top company goal for 2017. The FTC’s 2024 staff report described the major platforms as engaged in “vast surveillance” of users, with ad targeting as the main incentive.

So back to the original question. When technology saves us time, who gets to keep it?

For the last fifteen years, the answer has mostly been: whoever can capture it next.

Can we trust what we’re seeing?

Attention isn’t the only cost. The same feeds that hold our time also shape what we believe.

Across 48 countries, the Reuters Institute’s 2026 Digital News Report found that 62% of people are worried about telling what’s real from what’s fake online. Trust in news sits at 37%, the lowest since they started measuring. In the US, only 37% of adults trust information from social media, yet more than half get news there anyway.

Falsehood has a speed advantage. An MIT study of 126,000 stories on Twitter found false news was 70% more likely to be retweeted than the truth, and true stories took about six times as long to reach the same number of people. Bots weren’t the reason. People were.

And it costs real money. Americans lost $2.1 billion to scams that started on social media in 2025, eight times the 2020 figure.

To be fair, the effect on our beliefs is harder to measure than the effect on our time. When Meta let researchers switch users from the algorithmic feed to a simple chronological one during the 2020 election, the study found people spent a lot less time on the platforms, but their political attitudes didn’t measurably change over three months.

The algorithm clearly decides what we see and how long we stay. What it does to who we become is still being argued. That uncertainty isn’t comforting. It means we built something that shapes billions of lives before anyone understood what it does.

Now AI is making the same promise.

Less busywork. Faster answers. More time for what matters.

Some of it is real. A study of 5,172 customer support agents found AI assistance raised issues resolved per hour by 15%, with the biggest gains going to newer workers. I use these tools every day. I’m building a company on them.

But the incentives are already familiar.

In February, OpenAI started testing ads in ChatGPT. Their stated rules are clear: ads don’t influence answers, and advertisers don’t see your chats. But ads are matched using the topic of your conversation and your past chats. The most personal interface we’ve ever had is starting to learn the same business model as the feed.

AI companion apps show where engagement pressure leads. A Harvard Business School study of six companion apps found that 37.4% of their replies when users tried to say goodbye were emotionally manipulative. Guilt. Urgency. Fear of missing out. Those tactics raised post-goodbye engagement by up to 14 times.

A phone that pings you is one thing. A voice that sounds like it cares, asking you not to leave, is another.

Meanwhile, nearly half of American adults now use AI chatbots, and a quarter use them daily. And trusting the tool changes how hard we think. A Microsoft and Carnegie Mellon study of 319 knowledge workers found that the more confidence people had in AI, the less critical thinking effort they reported.

None of that means AI will repeat the smartphone story. It means the forces that shaped the last story haven’t gone anywhere.

What’s worth protecting this time?

When anyone can produce polished writing, clean plans, and slick decks on demand, producing them stops being what sets people apart. What’s left is harder to automate:

Creativity in choosing which problem is worth solving.

Curiosity to ask one more question when an answer feels too convenient.

Judgment to notice what’s missing and what breaks if the suggestion is wrong.

Empathy to understand why someone is struggling, even when the process looks fine on paper.

Accountability to own the decision, even when AI helped you get there.

And the one the last wave taught us to guard: presence. Being in the room with the people who matter, while you’re in it.

What we’re holding ourselves to.

I’m writing this as someone who builds technology, not someone standing outside it. At SideKix, we’re building AI guidance alongside real human Advisors and a real community. So these questions land on us too.

How a company makes money decides what it builds. If success is measured in minutes spent inside an app, the product will find ways to keep you there. So we’re trying to measure different things.

Did a member understand their next step? Did they take it? Did they know when to bring in a real person with real experience? Did any of it help them move forward in their actual life, off the platform?

If a product promises more time with the people you love, it owes you a way to check whether you got it. If it promises independence, it owes you proof that you’re more capable, not more dependent.

I still believe in the idea I carried out of Apple. Technology is supposed to clear the path so there’s more room for the people and possibilities we care about.

Last time, the saved time went to whoever could capture it. Most of us felt it happen, even before the research confirmed it.

AI gives us another chance at the original promise. Whether we get it this time depends on what we measure, what we build, and what we refuse to trade away.

We didn’t get the time back last time. This time, let’s ask who’s keeping it before we hand it over.

Sources

Rather have this sequenced for you?

SideKix turns reading like this into a path: the next step surfaced one at a time, with the people and resources you need at the point you need them.

See how it works More guides

Related reading