What Does ‘Dopagaki’ Mean? Short Videos, Algorithms, and the Theft of Attention

You meant to watch one short video.

Before you have formed a thought about it, your thumb has summoned the next. Humor, anger, surprise, fear: as soon as one stimulus fades, another arrives. You appear to be choosing what to watch, yet afterward you may struggle to say what you wanted in the first place.

Japanese internet culture has given this condition an ugly name: dopagaki.

The word is cruel. Its second half, gaki, is a contemptuous word for a child. It turns children and distracted users into immature people who supposedly lack discipline. But the discomfort the word captures should not be dismissed with the insult.

The better question is not who is a dopagaki? It is who benefits from making all of us more likely to behave this way?

The conclusion first: the problem is not brevity but optimization against exit

Short video is not inherently bad. A short clip can teach a skill, create joy, open a route into a distant political issue, or make knowledge accessible to someone who cannot begin with a long lecture.

The danger appears when four things are joined:

1. Time spent and advertisements viewed become revenue.
2. Infinite scroll, autoplay, and notifications remove natural stopping points.
3. Watch time, pauses, replays, comments, and shares become signals for selecting the next stimulus.
4. Billions of responses continuously improve the sequence most likely to prevent exit.

In such a system, a company does not necessarily need to learn what you reflectively value. A video you watch in anger, a rumor you dislike but repeatedly check, and a frightening post you cannot leave all register as engagement.

The fact that you watched something is not proof that you wanted it.

Individual restraint matters. But it is unfair to treat a tired person at the end of the day and a company equipped with enormous datasets, experiments, and behavioral design as equal decision-makers. Personal responsibility cannot settle the case.

What “dopagaki” means—and where it came from

The term is usually explained as a contraction of dōpamin chūdoku no gaki: roughly, “a dopamine-addicted brat.” It mocks, or sometimes self-mockingly describes, a person accustomed to rapid rewards from short videos, games, and gacha mechanics who struggles to wait or remain with slower material.

It is not a term in medicine, psychology, or education. It has no agreed age threshold, symptom list, or diagnostic test. The same person may scroll compulsively on a train and then concentrate for hours at work. It is more accurate to treat it as a name for behavior that can emerge in a particular environment, not a type of human being.

The current use is often traced to an X post from late December 2024. Discussing Mrs. GREEN APPLE’s song “Lilac,” the post used the term while suggesting that the music regularly introduced changes to prevent a stimulation-hungry audience from getting bored.

That does not prove a single inventor. Internet slang grows through quotation, alteration, and convergent coinage. The post is best understood as an important point of diffusion, not a birth certificate for the word.

Is “Lilac” a dopagaki song?

Music given the label is often said to have:

• a very short introduction or an immediate hook;
• changes in timbre, rhythm, vocal style, or harmony every few seconds;
• an early chorus or memorable line;
• many moments that work as isolated social-video clips;
• high event density and little empty space.

That is neither a diagnosis nor a musical genre. Rapid change long predates TikTok, and density can be an artistic technique rather than evidence of a damaged listener.

A musical analysis discussed by KAI-YOU argues that “Lilac” contains many small devices but also repeats motifs and follows an intelligible J-pop structure. Describing it as a random injection of novelty overlooks relations within the composition.

Phrases such as “asa no hikari no naka de dopagaki,” “dopagaki-only BGM,” Vocaloid-style spinoffs, and music edits are later meme play, not the term’s original source. “Dopagaki music” can be a provocative metaphor for contemporary editing, but it cannot determine a song’s quality or a listener’s intelligence.

“Adogaki” versus “dopagaki”: unstable slang is not a brain taxonomy

Two incompatible explanations of the related word adogaki circulate online:

• “an adrenaline-addicted brat,” seeking conflict and excitement;
• “an advice-giving brat,” offering unsolicited guidance to establish superiority.

There is no settled scientific definition choosing between them. Neatly classifying “dopagaki” as pleasure-seeking and “adogaki” as arousal-seeking may be memorable, but it is not a typology derived from neurological testing.

The instability is evidence of what these terms are: improvised criticism and insult, not medicine. Search-result definitions should not become personality tests for children.

There is no dopagaki diagnosis—and it is not ADHD

Watching films at double speed, reaching for a phone while waiting, or finding a long article difficult cannot by itself diagnose a disease or disorder.

As the U.S. National Institute of Mental Health explains, ADHD is a neurodevelopmental disorder involving persistent inattention, hyperactivity, or impulsivity that begins in childhood, appears in more than one setting, and interferes with functioning. Sleep deprivation, anxiety, depression, stress, and other conditions can produce similar difficulties, so professional assessment considers history and context.

“Dopagaki” has no agreed duration, setting, or impairment criteria. A person with ADHD may use short video; a person without ADHD may find it difficult to stop. Equating the two replaces a developmental condition with an insult and risks reducing needed support to “put down your phone.”

Online “dopagaki tests” are therefore not diagnostic. The following questions can instead support observation:

• Can I stop at the time I chose before opening the app?
• Do I feel satisfied afterward, or merely continue searching for a stronger stimulus?
• Is use repeatedly displacing sleep, study, work, or relationships?
• Would I actively seek this content if the recommendation feed disappeared?
• Is the movement of my thumb increasingly defeating my own wish to stop?

These questions measure distance between intention and action; they do not assign a derogatory identity. Persistent impairment deserves an appropriate medical, psychological, or educational consultation, not a meme diagnosis.

Dopamine is not “pleasure juice”

The word encourages a simple story: every swipe floods the brain with a pleasure chemical until the brain breaks. Neuroscience is more complicated.

Wolfram Schultz’s review of dopamine reward-prediction-error signaling describes dopamine’s role in learning, motivation, and updating value when an outcome is better or worse than expected. It is not merely a liquid meter of pleasure.

Short-video feeds can be compelling not because every clip is pleasurable, but because the next outcome is uncertain and an unexpected hit sometimes arrives. The recommender learns from that history, reducing irrelevant misses while preserving novelty. The user and the system learn from one another.

“Dopamine made me do it” is therefore no more complete than “my finger moved, so the choice was entirely free.” Freedom can become stronger or weaker within a learned environment.

What the research shows—and what it does not

Concern about short-form video is evidence-based, but permanent destruction of attention after a few clips has not been demonstrated.

Study	What was observed	What it cannot establish
Yan et al. (2024)	Among 48 young adults, greater short-video addiction tendency correlated with lower self-control and an EEG index related to executive control	That video use caused the neural difference
Luo et al. (2025)	In a 72-person randomized study, 30 minutes of TikTok-style clips was associated with weaker proactive preparation for a following task	Long-term impairment or a result that applies to every user
Tam et al. (2024)	Across several experiments, frequent switching could increase rather than relieve boredom and reduce satisfaction, attention, and meaning	An identical result across every age group and context
Mona et al. (2026)	A systematic review of 23 studies found attention and self-control were the most frequently reported cognitive concerns	A definitive conclusion beyond geographically concentrated and methodologically limited evidence

Two cautions matter.

First, much evidence is correlational. People already experiencing attention difficulty or stress may use short video more, so causation may run in both directions.

Second, uncertainty is not evidence of harmlessness. A small effect repeated daily across hundreds of millions of people can become socially important. The responsible response is not panic but longitudinal research, independent audits, and researcher access to platform data.

Did I watch it because I liked it—or because it moved me?

A recommender cannot read a mind. It treats clicks, dwell time, completion, sharing, comments, and replay as proxies for preference.

But a proxy can diverge from a person’s judgment.

A 2025 audit of Twitter’s ranking system found that engagement ranking amplified emotionally charged and politically hostile posts. Participants’ explicitly stated preferences, however, did not align with the political posts selected by the algorithm.

This does not show that an algorithm brainwashed everyone. It establishes a crucial possibility:

What provokes a response and what a person wants are not necessarily the same.

A frightening clip, a post that triggers an angry rebuttal, or a rumor checked with disgust may all look like success to an optimization system. When a company says “users chose it,” it may not distinguish reflection, regret, happiness, or long-term value.

The fear of losing track of what one likes is not simply the fear of outside influence. Friends, families, advertising, and culture have always shaped taste.

What is new is the scale and speed of a system that continuously measures each person’s smallest responses and supplies an individualized next stimulus before reflection begins.

Herbert Simon: abundant information creates scarce attention

Economist and cognitive scientist Herbert Simon argued that an abundance of information consumes the attention of its recipients.

The scarce object in the short-video market is not video. Video is supplied in effectively unlimited quantities. What is scarce is the few seconds before a person chooses another action.

From this perspective, “you used a free service, so it is your responsibility” is incomplete. The user is a customer but also a supplier of attention. Closing the screen destroys a revenue opportunity for the platform.

This does not require every employee to intend harm. Metrics are sufficient. If watch time, daily activity, and consecutive views are rewarded as success, even a well-intentioned designer faces pressure to delay exit.

Harry Frankfurt: what we want and what we want to want

Harry Frankfurt did not reduce freedom to the ability to act on whatever desire is strongest at a moment.

We may have a first-order desire to watch one more clip. We may simultaneously have a higher-order will: “I want to sleep,” “I do not want this impulse to govern me,” or “I want to remain capable of enjoying a long book.”

If the swipe alone counts as consent, the immediate desire always wins. Yet freedom depends on whether the desire a person reflectively identifies with can guide action—not merely on which urge is currently strongest.

Infinite scroll expands the number of available choices. In that respect it can increase freedom. But when it removes the stopping points needed to execute a higher-order wish to stop, it weakens freedom in another respect. More options and stronger self-government are not the same thing.

Kant: autonomy is more than the absence of an order

For Kant, autonomy is not simply being left without external commands. It is the ability to examine principles and give oneself the law one follows.

A recommendation screen issues no explicit command. No one physically forces the next swipe. Yet a system can still cultivate heteronomy if it targets pre-reflective reactions and works through vulnerabilities the user cannot see.

The distinction between persuasion and manipulation matters:

• Persuasion presents reasons and preserves room to refuse and reconsider.
• Manipulation routes around deliberation and uses hidden weaknesses to direct conduct.

Not every recommendation is manipulation. But an unexplained system that obscures stopping and reset tools, or monetizes children’s still-developing capacity for impulse control, moves away from treating people as ends in themselves.

Bernard Stiegler: institutions cultivate attention; industry consumes it

Bernard Stiegler described a tension between families, education, and culture, which cultivate attention over long periods, and marketing systems, which capture it for immediate response.

Attention is not a fixed tank of fuel delivered at birth. Waiting, listening, rereading, tolerating boredom, and staying to the end of another person’s account are learned with others.

Attention takes years to cultivate and seconds to monetize. A school or family may slowly form a capacity that a platform can auction with one notification.

When “dopagaki” points only at children, this asymmetry disappears. Children did not design the market surrounding them. Adults, too, enter the same system through exhaustion, loneliness, and pressure at work.

Why personal-responsibility rhetoric is unjust

There is a truth inside the personal-responsibility argument. Other people cannot entirely perform the practice of putting down a device, disabling a notification, or building a stopping routine on someone’s behalf. Treating users as having no agency would be another insult.

But placing all responsibility there is unjust.

First, information is asymmetric. Platforms learn abandonment rates through large experiments; users do not know which experiments were performed on them.

Second, capacity is asymmetric. Children and people under sleep loss, loneliness, or stress may be more affected by the same design.

Third, profit is asymmetric. Failed self-control costs the user sleep, learning, and relationships, while the added watch time earns money elsewhere.

Fourth, formal choice can conceal engineered difficulty. An exit button does not absolve a system when infinite supply, autoplay, notifications, and personalization are integrated to delay its use.

The U.S. Surgeon General’s advisory on youth social media likewise rejects placing the burden only on children and families. It assigns duties to platforms and policymakers, including safety assessment, data access, and age-appropriate design.

Responsibility is not all or nothing. The actor with greater knowledge, design power, profit, and ability to prevent harm should carry more of it.

What meaningful brakes would look like

A 2024 Federal Trade Commission report identifies features such as autoplay, infinite scroll, notifications, popularity metrics, and data-driven recommendation as tools that can extend engagement.

In 2026 the European Commission issued preliminary findings—not final judgments—concerning potentially addictive design at TikTok and later Instagram and Facebook. It assessed infinite scroll, autoplay, notifications, and highly personalized recommendation as a system of interacting risks rather than isolated conveniences.

The answer is neither a total ban on short video nor more shame for users.

1. Restore stopping points by default

After a finite batch, require a meaningful action to load more. Make nighttime stops and autoplay-off defaults for minors, not settings buried several menus deep.

2. Distinguish response from preference

Measure later satisfaction, regret, hiding, and “show less of this” choices—not only dwell time. Give explicitly followed topics greater weight than an involuntary pause.

3. Make recommendation understandable and breakable

Explain why a particular item appeared. Provide readable histories, partial deletion, full reset, and obvious chronological or non-personalized feeds.

4. Protect minors’ attention from raw engagement targets

Use age-appropriate defaults, stop nighttime notifications, limit consecutive viewing, and offer non-profiled use. Reliance on parental surveillance alone turns household inequality into safety inequality.

5. Permit independent verification

Promises that a platform “cares about wellbeing” are not enough. Researchers need access to appropriate data, and companies should publish risk assessments, feature-change evaluations, and audits of effects on minors.

6. Treat individual countermeasures as first aid

Disable notifications. Remove the app from the home screen. Choose an end time before starting. Use following or chronological views. Keep the phone outside the bedroom. Practice staying with one long work each day.

These measures can help, but they are first aid in a polluted environment. Teaching people to wear a mask does not erase the emitter’s responsibility.

The objection: did television and advertising not shape taste too?

Yes. Preferences have never emerged from a pure, untouched interior. Popular music, television schedules, bookstore shelves, friends, and advertising have always shaped what people know and want.

There was no golden age of completely autonomous taste before algorithms.

The differences are nonetheless substantial:

• individual response is measured by the second;
• the supply has no endpoint;
• the same system both measures behavior and selects the next stimulus;
• the influencing party learns more about what worked than the influenced person does;
• experiments and optimization can be repeated across millions of people at high speed.

The problem is not influence itself. It is influence that is opaque, costly to refuse, and continuously improves for only one side.

Turn the word upward

“Dopagaki” is useful and dangerous.

It is useful because it compresses a recognizable experience: becoming accustomed to switching stimuli and finding it harder to remain with slower time.

It is dangerous because it attaches the outcome to a child’s character while allowing the designers, investors, and institutions to disappear beyond the screen.

If the word is to be used, it should not be aimed downward.

It should name a market in which whoever captures attention wins and the person whose attention was captured pays the cost of recovering it.

Freedom cannot mean freedom from all influence; that is impossible. It means being able to discover what influenced us, stop, and choose again what we want to want. A design that systematically removes the time, explanation, and stopping points required for that revision is no longer merely a convenience issue. It is a political and ethical one.

We do not need to condemn people who enjoy short videos. The demand is more modest and more serious: they must be able to leave, refuse a recommendation, and reconstruct their own taste. Responsibility for protecting that capacity belongs not only to individuals but especially to those who know the most, design the system, and profit from it.

---

NOW IN QUESTION

Was the thing you watched for a long time yesterday something you still wanted today?

Or was your failure to leave merely renamed a preference?

---

References

• KAI-YOU, “Is Mrs. GREEN APPLE’s ‘Lilac’ really music for dopagaki?” (Japanese)
• Wolfram Schultz, “Dopamine reward prediction-error signalling,” Nature Reviews Neuroscience (2016)
• Yan et al., “The effect of short-form video addiction on users’ attention,” Frontiers in Human Neuroscience (2024)
• Luo et al., “Context- and Time-Dependent Cognitive Effects of Short-Form Video Use,” Behavioral Sciences (2025)
• Tam et al., “Fast-Forward to Boredom,” Journal of Experimental Psychology: General (2024)
• Mona et al., “The relationship between short-form video use and cognitive functioning,” Journal of Youth Studies (2026)
• Milli et al., “Engagement, user satisfaction, and the amplification of divisive content on social media,” PNAS Nexus (2025)
• U.S. Surgeon General, “Social Media and Youth Mental Health” (2023)
• Federal Trade Commission, “A Look Behind the Screens” (2024)
• NIMH, “Attention-Deficit/Hyperactivity Disorder: What You Need to Know”
• European Commission, “Preliminary findings on TikTok’s addictive design” (2026)
• Harry Frankfurt’s account of higher-order desires, Stanford Encyclopedia of Philosophy
• Bernard Stiegler, Taking Care of Youth and the Generations