Two hours and forty-three minutes. That's how long the average person spends on social media every single day in 2025. Not the doom-scrolling teen stereotype. You. Me. Everyone. That's more time than most people spend exercising in a week, and it didn't happen by accident.
I have been daily-driving this problem for a week, except the "product" isn't a gadget. It's my own feed. I downloaded screen time trackers, read the research, and sat with the uncomfortable reality that the apps I open reflexively every morning were built to make sure I never stop opening them. What I found was not what Devon would call an engineering marvel or what Audrey would call late capitalism's greatest heist. It was simpler and more annoying than both: a very good machine optimized to make me a worse version of myself, invisibly, one scroll at a time.
The Machine Is Not Neutral
Here is what social media algorithms actually do in 2025. They are not just sorting content you asked for. They are predicting who you're about to become and feeding you the content that will make that prediction come true. Meta's system now runs what it calls "intent modeling," forecasting what you'll want to engage with based on cross-platform behavior pulled from Facebook, Instagram, and Threads simultaneously. TikTok's For You feed uses real-time behavioral signals including watch time, pauses, replays, and swipe patterns, not just what you liked but how long you hovered before moving on. These systems use deep and reinforcement learning models trained on massive datasets, tracking likes, comments, shares, watch time, and even pauses, all to rank content that maximizes your engagement.
A 2025 study out of ScienceDirect put this to a direct test with 88 TikTok users. When their feeds were switched from highly personalized to less personalized for just one week, daily usage frequency dropped, self-regulation improved, and enjoyment actually decreased. That last part is the tell. Users spent less time but liked themselves more for it. The algorithm's job isn't to make you happy. It's to make you return. Those are not the same goal.
Meanwhile, a NATO Strategic Communications Centre of Excellence experiment found that for just €10, a bad actor can generate hundreds of videos and thousands of images using AI tools. On X, 156,000 fake views now cost €10. The barrier to large-scale behavioral manipulation has never been lower. Research from Concordia University confirmed that simple reinforcement-learning bots, given only two pieces of information about you, your current opinions and your follower count, can efficiently maximize polarization across a social network. The algorithm built to keep you engaged and the algorithm built to radicalize you are running the same basic playbook.
What It's Actually Doing to Your Brain
Is it worth it? Let us talk numbers, because the numbers are bad.
Around 210 million people worldwide are estimated to be addicted to social media. In the United States, approximately 40% of Americans aged 18 to 22 acknowledge their own addiction. Users who spend more than three hours a day on social media are 2.6 times more likely to show signs of depression. Among 18-to-29-year-olds, 41% report feeling anxious when they haven't checked their feeds in a few hours. Roughly 35% of users now spend more time on algorithm-driven discovery feeds than on content from people they actually chose to follow.
This is not a bug. Published research in a peer-reviewed neuroscience journal put it plainly: AI-driven social media algorithms are designed to maximize screen time, deepening activation of the brain's reward centers, and this cycle of optimized content accelerates the development of addictive behaviors. Infinite scrolling, personalized notifications, and algorithm-driven content deliver dopamine hits through intermittent rewards. The mechanism is functionally identical to a slot machine. The platforms just dressed it in a friendlier UI.
Teens are absorbing the worst of it. Fifty-one percent of U.S. teenagers spend 4.8 hours daily on social media. Girls between 11 and 15 told the Surgeon General's office they felt addicted to specific platforms. The TikTok algorithm can encourage addictive behavior by promoting excessive user engagement, which can negatively affect mental health, decrease attention spans, and disrupt daily activities. Utah sued Snapchat in July 2025 for designing features specifically to addict children. That's not a fringe legal theory anymore. That's a state government saying out loud what the data has been showing for years.
The Verdict on What You Should Actually Do
I'm not going to tell you to delete your apps. That's not realistic advice for most people, and frankly, some of these platforms do deliver genuine value. TikTok finds me products I actually buy. Instagram keeps me connected to people I genuinely care about. I'm not pretending those things are worthless.
But I am going to tell you that the algorithm is not your friend, and knowing that is the starting point. A 2025 cross-sectional survey of 1,200 users across X, TikTok, YouTube, and Facebook found that users who understand how algorithms work report higher trust, more intentional use, and better overall outcomes from their social media experience. Algorithm awareness correlated positively with every good behavioral outcome they measured. Knowing the machine exists and knowing what it wants from you changes how you interact with it.
So: set a screen time limit and actually honor it. Get off the discovery feed and go back to following real people you like. If a piece of content is making you furious or miserable, recognize that the algorithm surfaced it on purpose, because outrage keeps you scrolling. Social networks are vulnerable to AI-driven manipulation and polarization by design, because their design fosters echo chambers by connecting users with like-minded content. You are not immune to this. Neither am I.
My wallet hates me but my setup is fire. Except in this case, the setup costs you nothing except the attention span and emotional equilibrium you've been quietly handing over every day. That's a worse deal than any gadget I've ever bought. At least a $1,200 phone shows up at your door. The algorithm just takes.
