EvilEye Amulet
YouTube Addiction

YouTube's algorithm decides what you watch. 70% of the time, you never chose it.

YouTube's own Chief Product Officer admitted that 70% of all watch time comes from algorithmic recommendations. You open the app to watch one video. The algorithm decides you will watch twelve. This is not a bug. It is the business model.

Escape the Rabbit Hole
5.0 on the App Store
70%
of all YouTube watch time comes from algorithmic recommendations, not your intentional choices
The Numbers

YouTube by the numbers: the world's deepest rabbit hole.

YouTube is the second-most visited website on Earth. Over 2.5 billion people use it every month. And its algorithm has been quietly optimized for one thing above all else: keeping you watching.

0 min
Average daily YouTube use worldwide
Backlinko, 2025
0B
Monthly active YouTube users globally
Business of Apps, 2025
0B hrs
Hours of video watched on YouTube every single day
YouTube Official Blog
0%
Of watch time driven by algorithmic recommendations
YouTube CPO, CES 2018; Quartz

From video sharing to attention-maximizing machine

2005

YouTube launches as a video-sharing site

"Broadcast Yourself" was the original tagline. Users uploaded and shared videos. You searched for what you wanted, watched it, and left. The platform was a tool, not a trap.

2012

The algorithm shifts from views to watch time

YouTube changes its recommendation algorithm to prioritize watch time over view count. This single decision transforms the platform from rewarding what people click on to rewarding what keeps them watching longest. Content creators begin optimizing for retention, not quality.

2015

Autoplay becomes the default

YouTube enables autoplay by default, automatically starting the next recommended video when the current one ends. The decision to keep watching is removed. Continuing becomes the default. Stopping requires active intervention.

2019

$170 million FTC fine for tracking children

Google and YouTube pay a record $170 million settlement for violating the Children's Online Privacy Protection Act. The FTC found YouTube had been tracking viewing data from children under 13 without parental consent, using it to serve targeted ads on child-directed content.

2020 – 2021

YouTube Shorts launches, adding a second addiction layer

YouTube launches Shorts to compete with TikTok, adding an infinite-scroll short-form video feed to its already addictive long-form platform. By 2024, Shorts reaches 70 billion daily views. Users now have two distinct compulsion loops in a single app.

2024

Kids Online Safety Act passes the Senate 91–3

The US Senate overwhelmingly passes the Kids Online Safety and Privacy Act, specifically targeting platforms like YouTube that exploit youth vulnerabilities. YouTube reaches 2.5 billion monthly active users while researchers continue documenting the harms of its recommendation system.

The Algorithm

How YouTube hooks you.

YouTube is not just a video platform. It is a watch-time maximization engine built on autoplay, parasocial bonds, and an algorithm that knows what keeps you watching better than you do.

The Recommendation Algorithm: Your Choices Are Not Your Own

YouTube's Chief Product Officer stated publicly that over 70% of all watch time on the platform comes from algorithmic recommendations. That means the vast majority of what you watch on YouTube was not chosen by you — it was chosen by an algorithm optimized for one metric: total watch time. The system does not recommend what is best for you, most educational, or most satisfying. It recommends whatever keeps you watching longest. Your viewing history, watch duration, click patterns, and thousands of other signals feed a deep learning model that builds a continuously updating profile of your attention vulnerabilities.

YouTube CPO at CES 2018; Quartz, "YouTube's recommendations drive 70% of what we watch" (2018); Hootsuite (2025)

Autoplay and Up Next: The Rabbit Hole Engine

Autoplay is the mechanism that converts a single video into a multi-hour session. When your video ends, the next one starts automatically — no decision required. Research from Penn State University describes this as "choice architecture" that exploits path-of-least-resistance psychology: continuing requires less effort than stopping. A study published in the International Journal of Human-Computer Studies found that autoplay leads users progressively deeper into increasingly narrow content. The researchers call this the "rabbit hole effect" — a collapse of mainstream recommendations into ultra-personalized feeds that lock users into specialized loops.

Penn State University (2024); ScienceDirect, "Preventing users from going down rabbit holes" (2024)

YouTube Shorts: TikTok's Dopamine Layer, Added to an Already Addictive Platform

In 2020, YouTube launched Shorts — an infinite-scroll, full-screen, short-form video feed virtually identical to TikTok. By 2024, Shorts had reached 70 billion daily views, up from 30 billion in 2021. Dr. Anna Lembke, chief of Stanford University's dual diagnosis addiction clinic, has stated that brief attention-grabbing videos act as "powerful stimuli triggering dopamine surges akin to other addictive behaviors." What makes YouTube uniquely dangerous is that Shorts is not a standalone app — it is layered on top of an already addictive long-form platform. Users now face two distinct compulsion mechanisms in a single app: the autoplay rabbit hole and the infinite scroll dopamine feed.

Dr. Anna Lembke, Stanford University; Awisee, YouTube Shorts Statistics (2025); DemandSage (2025)

Parasocial Relationships: You Think They Are Your Friends

Unlike TikTok or Instagram, YouTube fosters deep parasocial relationships — one-sided psychological bonds where viewers feel genuine emotional connection to creators they have never met. Research published in the Journal of Social Media Studies found that individuals with greater YouTube exposure develop stronger parasocial relationships and place higher relational importance on these bonds. Studies show these parasocial ties satisfy emotional, behavioral, and cognitive needs similarly to real friendships: they reduce loneliness, influence decisions, and viewers even assimilate creators' traits. The danger is that these bonds create a compulsion to return to YouTube not for content, but to "visit friends" — making it emotionally costly to reduce usage.

Journal of Social Media Studies (2019); Scientific Reports, "Parasocial relationships perceived as effective at fulfilling emotional needs" (2024); ResearchGate (2022)
YouTube touted its popularity with children to prospective corporate clients. Yet when it came to complying with COPPA, the company refused to acknowledge that portions of its platform were clearly directed to kids.
— FTC Chairman Joe Simons, announcing the record $170 million YouTube settlement (September 2019)
The Research

What the research says about YouTube and your brain.

This is not speculation. These findings come from peer-reviewed studies, systematic reviews, and YouTube's own data.

Sleep Is Being Stolen by Autoplay

A 2025 study published in JAMA Network Open found that people who used screens before bed had a 33% higher rate of poor sleep quality and slept approximately 50 minutes less each week. More critically, each one-hour increase of screen time after going to bed was tied to a 59% higher chance of insomnia symptoms. YouTube's autoplay feature is particularly damaging for sleep because it removes the natural stopping point that would allow a viewer to put down their phone. YouTube itself acknowledged this problem by launching a sleep timer in August 2024 for Premium subscribers, later expanding it to free users — an implicit admission that its own product keeps people watching past bedtime.

JAMA Network Open (2025); Sleep Education, "Screen time and sleep: What new studies reveal" (2025); YouTube Sleep Timer launch (2024)
59%
higher chance of insomnia symptoms for each extra hour of screen time in bed

The Rabbit Hole Pulls You Toward Extremes

A systematic review examining 23 studies on YouTube's recommendation system found that 14 studies implicated the algorithm in facilitating problematic content pathways, seven produced mixed results, and only two did not implicate the system at all. Research from UC Davis found that for certain users, video recommendations are more likely to come from channels that share political extremism, conspiracy theories, and otherwise problematic content. The rabbit hole effect works because each recommended video is slightly more engaging or sensational than the last, keeping the viewer watching while gradually shifting the content toward increasingly narrow or extreme territory. A Penn State study showed that giving users control over autoplay helps them recognize when they are going down a rabbit hole — but YouTube keeps autoplay enabled by default.

PMC, "Systematic review: YouTube recommendations and problematic content" (2021); UC Davis (2022); Penn State University (2024)
14/23
studies implicated YouTube's algorithm in facilitating problematic content pathways

Compulsive Watching Follows Addiction Patterns

A 2024 study published in the International Journal of Mental Health and Addiction developed and validated the YouTube Addiction Scale, based on the established component model of addiction. The research found that compulsive YouTube use mirrors the hallmarks of behavioral addiction: salience (YouTube dominates your thoughts), mood modification (using it to regulate emotions), tolerance (needing more to get the same effect), withdrawal (irritability when you cannot access it), conflict (it interferes with relationships and responsibilities), and relapse (failed attempts to cut back). While YouTube addiction is not yet an official DSM-5 diagnosis, researchers argue it meets the clinical criteria for a behavioral addiction, with binge-watching associated with increased depression, anxiety, stress, and social isolation.

International Journal of Mental Health and Addiction, "The YouTube Addiction Scale" (2024); PMC, "Unveiling the YouTube addiction" (2024)
6
components of addiction identified in compulsive YouTube use: salience, mood modification, tolerance, withdrawal, conflict, relapse

Dopamine Loops Drive Binge Watching

YouTube offers a constant stream of video content that, over time, becomes precisely targeted to an individual's preferences, creating what addiction researchers describe as a cycle of small, repetitive dopamine hits. Videos deliver inconsistent satisfaction levels through variable reward mechanisms — the same unpredictable reinforcement schedule that makes slot machines compulsive. Research has found that uncontrolled and repetitive social media use impacts the brain in the areas that govern impulse control and reward processing. The combination of variable reward with autoplay creates a particularly potent loop: each video might be the best one, and the next one starts before you can decide to stop. More than 1 in 5 YouTube users report watching for five or more hours per day.

Birches Health, "YouTube Social Media Addiction" (2024); Canadian Centre for Addictions; AddictionCenter.com
21.7%
of YouTube users report watching for 5 or more hours every single day

Parasocial Bonds Make Quitting Feel Like Losing a Friend

Research on parasocial relationships reveals why YouTube is uniquely difficult to quit. A 2024 study published in Scientific Reports found that participants felt their YouTube creator parasocial relationships were more effective at fulfilling emotional needs than in-person acquaintances. People form one-sided emotional bonds with creators who share personal stories, daily routines, and intimate thoughts through vlogs. These bonds satisfy real psychological needs — reducing loneliness, providing companionship, and creating a sense of belonging. The consequence is that reducing YouTube use does not just feel like giving up entertainment. It feels like abandoning relationships. This emotional dimension sets YouTube apart from short-form platforms where the connection is to the feed, not to specific people.

Scientific Reports, "Parasocial relationships perceived as effective at fulfilling emotional needs" (2024); ScienceDirect, "Parasocial relationships and YouTube addiction" (2022)
1-sided
Parasocial bonds feel like real friendships but the creator does not know you exist
Platform Comparison

YouTube is a different kind of trap. It has two.

Other platforms have one addiction mechanism. YouTube has two: long-form autoplay rabbit holes and short-form infinite scroll. Here is how daily usage compares across platforms.

TikTok 95 min/day
YouTube 49 min/day
Instagram 33 min/day
Facebook 31 min/day
Snapchat 30 min/day
X (Twitter) 24 min/day
Sources: Backlinko 2025, Statista 2023, DataReportal 2024

Why YouTube's dual-format approach is uniquely dangerous

Long-Form

Autoplay Rabbit Holes

Traditional YouTube videos create deep engagement through extended watch sessions. The autoplay system chains videos together, each one recommended by an algorithm optimized for watch time. You start watching a 10-minute tutorial and surface 90 minutes later having watched seven increasingly tangential videos. The content is long enough to create parasocial bonds with creators, making returning feel like visiting friends. This mechanism is unique to YouTube — no other major platform combines algorithmic recommendations with multi-minute content and autoplay at this scale.

Shorts

Infinite Scroll Dopamine Feed

YouTube Shorts replicates TikTok's addiction mechanics: full-screen immersion, infinite vertical scroll, and algorithmic curation. With 70 billion daily views in 2024, Shorts adds a rapid-fire dopamine layer on top of the existing autoplay system. Users who might have closed YouTube after their video ended now get funneled into Shorts. The transition between formats is seamless and often automatic. The result is a platform with no escape hatch — long-form hooks you with depth and parasocial bonds, Shorts hooks you with speed and novelty. Both feed the same watch time metric.

Brief attention-grabbing videos act as powerful stimuli triggering dopamine surges akin to other addictive behaviors. The rapid and easily consumable nature of short-form videos can elicit high levels of dopamine.
— Dr. Anna Lembke, Chief of Stanford University's Dual Diagnosis Addiction Clinic
The Youth Crisis

For children, YouTube is a documented danger.

From the $170 million COPPA fine to the Elsagate scandal to the Kids Online Safety Act — YouTube has a long, documented history of failing to protect its youngest and most vulnerable users.

$170M

Record FTC Fine for Tracking Children

In September 2019, Google and YouTube paid a record $170 million settlement for violating COPPA by collecting personal data from children under 13 without parental consent. YouTube had been tracking viewing history on child-directed channels to serve targeted ads. The FTC noted YouTube told toy companies like Mattel it was "today's leader in reaching children age 6-11."

Federal Trade Commission Press Release, September 2019
93%

Of US Teens Use YouTube

Pew Research found that 93% of US teens use YouTube, making it the most-used platform among adolescents. 71% visit daily and 16% report being on it "almost constantly." 49% of teen girls in the US reported feeling addicted to YouTube or using it for longer than they intended. Young adults spend an average of 77 minutes per day on the platform.

Pew Research Center, 2024; ElectroIQ, 2025
91–3

Senate Vote for Kids Online Safety Act

In July 2024, the US Senate passed the Kids Online Safety and Privacy Act by a near-unanimous 91 to 3 vote. The bill prohibits targeted advertising to minors, requires platforms to acknowledge and protect children, and prevents companies from exploiting youth vulnerabilities for profit. YouTube is one of the primary targets of this legislation.

US Senate, Kids Online Safety and Privacy Act (2024)

The Elsagate scandal and algorithm-generated content

Disturbing Content Targeting Children

In 2017, the Elsagate scandal exposed a massive network of bizarre and disturbing videos on YouTube and YouTube Kids, featuring popular children's characters like Elsa, Spider-Man, and Peppa Pig in violent, sexualized, or traumatizing scenarios. These videos were algorithmically promoted to children because they generated enormous watch time. The content was often auto-generated at scale, optimized not for child safety but for the algorithm's watch time metric. The scandal revealed that YouTube's recommendation system, when left unchecked, will promote whatever retains attention — regardless of who is watching or what the content contains.

ChildHub, "The Elsagate situation" (2017); Multiple news investigations (2017-2018)

YouTube Kids Still Fails to Protect

Despite launching YouTube Kids as a "safer" alternative, research continues to document problems. A 2024 study on YouTube Kids found that videos with "negative interactions" are impacting children's mental health. A separate study published in PubMed assessed the educational value of YouTube Kids videos related to anxiety, depression, and ADHD, finding significant quality concerns. The fundamental problem remains: YouTube Kids relies on the same algorithmic recommendation system that prioritizes engagement over safety. YouTube told Hasbro it was the "#1 website regularly visited by kids" while simultaneously claiming its platform was not directed at children.

Tubefilter, "YouTube Kids videos impacting mental health" (2025); PubMed (2024); FTC Settlement Documents (2019)
YouTube told toy manufacturers that YouTube is today's leader in reaching children age 6-11 against top TV channels. Yet when it came to complying with the law, the company refused to acknowledge that portions of its platform were clearly directed to kids.
— FTC Chairman Joe Simons, September 2019
The Solution

Breaking the YouTube rabbit hole does not require willpower.

YouTube's entire design is built to eliminate stopping cues — autoplay removes the decision to continue, Shorts removes the decision to scroll, and parasocial bonds remove the desire to leave. The solution is to put a stopping cue back before the first video even plays.

YouTube's addiction works because of three compounding mechanisms: frictionless access (tap and you are watching), no stopping cues (autoplay chains videos endlessly), and emotional attachment (parasocial bonds make leaving feel like abandoning friends). EvilEye targets the first mechanism. By introducing a brief, intentional pause before YouTube opens, it interrupts the automatic reflex that feeds the entire system.

Penn State researchers found that giving users control over autoplay helps them recognize when they are going down a rabbit hole. EvilEye goes further: it intervenes before the rabbit hole begins. The smile requirement shifts you from reactive autopilot to conscious decision-making. Instead of "I just opened YouTube without thinking," you arrive at "I am choosing to open YouTube right now, and I am choosing how long I want to spend." That shift changes everything.

1

Smile to Interrupt

When you reach for YouTube on autopilot, EvilEye catches you. Before the app opens, it asks for a genuine smile using your iPhone's TrueDepth camera. This two-second pause is the stopping cue that YouTube deliberately removed. You shift from reacting to choosing — before the first video plays, before autoplay starts, before the rabbit hole begins.

2

Choose Your Time

After smiling, you decide how long you want YouTube unlocked. Ten minutes to watch a specific video? Thirty minutes to catch up on a creator? The choice is yours. The critical difference is that you set a boundary before the algorithm has a chance to pull you deeper. YouTube removes stopping cues. EvilEye puts one at the very beginning.

3

Stay Protected

When your chosen time expires, EvilEye steps back in. No willpower drain. No internal negotiation against an algorithm designed by thousands of engineers. The app locks again and the rabbit hole closes. Over time, the number of times you reflexively reach for YouTube decreases — because your brain learns there is friction waiting at the door.

Download EvilEye Free
FAQ

YouTube addiction: your questions answered.

Yes. While YouTube addiction is not yet recognized as an official diagnosis in the DSM-5, researchers have developed validated instruments like the YouTube Addiction Scale based on the component model of addiction. Studies show YouTube uses the same psychological mechanisms as other addictive platforms: variable reward schedules, autoplay that eliminates stopping cues, and algorithmic recommendations that drive over 70% of all watch time. 93% of US teens use YouTube, with 71% visiting daily and 16% reporting they are on it constantly. More than 1 in 5 users say they watch for 5 or more hours per day.
YouTube's recommendation algorithm drives over 70% of all watch time on the platform, according to YouTube's own Chief Product Officer. The algorithm is optimized for watch time, not user satisfaction, meaning it recommends whatever keeps you watching longest rather than what is most valuable to you. Combined with autoplay, which automatically starts the next video without requiring any decision from the viewer, this creates a frictionless rabbit hole effect. Research from Penn State University has shown that autoplay exploits path-of-least-resistance psychology, where continuing requires less effort than stopping.
YouTube rabbit holes occur when the recommendation algorithm progressively leads viewers deeper into increasingly narrow or extreme content. A systematic review found that 14 out of 23 studies implicated YouTube's recommender system in facilitating problematic content pathways. The phenomenon works because each recommended video is slightly more engaging or extreme than the last, keeping the viewer watching while gradually shifting the content. Penn State research has shown that giving users control over autoplay can help them recognize when they are going down a rabbit hole, but YouTube keeps autoplay enabled by default.
YouTube Shorts uses essentially the same addiction mechanics as TikTok: infinite vertical scroll, full-screen immersion, and algorithmic curation. YouTube Shorts reached 70 billion daily views in 2024, up from 30 billion in 2021. Dr. Anna Lembke, chief of Stanford University's dual diagnosis addiction clinic, has stated that brief attention-grabbing short-form videos act as powerful stimuli triggering dopamine surges akin to other addictive behaviors. What makes YouTube uniquely dangerous is that Shorts adds a TikTok-style dopamine layer on top of an already addictive long-form platform, giving users two distinct addiction mechanisms in a single app.
Parasocial relationships are one-sided psychological bonds viewers form with creators they have never met. Research published in the Journal of Social Media Studies found that individuals with greater YouTube exposure develop stronger parasocial relationships with YouTube personalities and place higher relational importance on these bonds. Studies show these parasocial ties can satisfy emotional, behavioral, and cognitive needs similarly to real friendships, making viewers feel less lonely. The danger is that these bonds create a compulsion to return to YouTube not for content, but to visit perceived friends, making it feel emotionally costly to reduce usage.
YouTube has a documented history of harming children. In 2019, Google and YouTube paid a record $170 million FTC settlement for violating the Children's Online Privacy Protection Act by collecting data from minors under 13 without parental consent. The Elsagate scandal exposed algorithm-generated disturbing content targeting children through popular characters. 49% of teen girls in the US report feeling addicted to YouTube or using it longer than intended. In 2024, the US Senate passed the Kids Online Safety and Privacy Act by a 91 to 3 vote, specifically targeting platforms like YouTube that exploit youth vulnerabilities for profit.
Yes. A 2025 study published in JAMA Network Open found that people who used screens before bed had a 33% higher rate of poor sleep quality and slept about 50 minutes less each week. Each one-hour increase of screen time after going to bed was tied to a 59% higher chance of insomnia symptoms. YouTube's autoplay feature is particularly problematic for sleep because it removes the natural stopping point that would allow a viewer to put down their phone. YouTube itself acknowledged this issue by launching a sleep timer for Premium subscribers in 2024, later expanding it to free users.
Yes. EvilEye directly targets YouTube's core addiction mechanism: frictionless, automatic access. Before you can open YouTube, EvilEye requires a genuine smile using your iPhone's TrueDepth camera. This two-second pause interrupts the autopilot reflex that powers YouTube's rabbit hole effect. You then choose how long you want YouTube unlocked. This approach directly counters autoplay's design by reintroducing the stopping cue that YouTube deliberately removed. Instead of falling into an unplanned three-hour binge, you make a conscious decision about your time before the first video even plays.

Related Reading

TikTok Addiction The algorithm designed to keep you scrolling Netflix Addiction Auto-play and the binge-watching trap Twitch Addiction Gambling streams and parasocial bonds
EvilEye Amulet

YouTube was designed to keep you watching.
EvilEye was designed to give you the choice.

You now know how YouTube's algorithm works, how autoplay removes your stopping cues, and how parasocial bonds keep you coming back. The only question left is whether you will keep falling down the rabbit hole — or take conscious control before the next video plays.

Download EvilEye — It's Free
5.0 on the App Store