top of page

Psychological Encyclopedia

Dark Patterns: How Interface Design Shapes Digital Choices and Attention

2 days ago
29 min read

Author: Ukrainian Psychological Hub · Published: September 27, 2026 · Editorial Policy


Dark patterns are interface designs that steer, pressure, confuse, obstruct, or deceive people in ways that can make a platform-preferred choice easier than the choice a user would otherwise make. They can appear in subscription flows, privacy settings, shopping checkouts, app permissions, account deletion, notifications, advertising, games, and many other digital environments. The important psychological point is that choices are made through an interface, and an interface determines what is visible, what is emphasized, what is preselected, how much effort alternatives require, when information appears, and how often a request returns.


Research does not support the idea that dark patterns work through one mysterious “brain hack,” a single dopamine mechanism, or universal mind control. Their effects are better understood through ordinary processes of attention, perception, decision-making, learning, effort, emotion, and context. A 2026 systematic review of user experiments concluded that deceptive or manipulative interface patterns reliably altered behavior across the experimental literature, while effect sizes varied substantially by pattern and setting. It also found that many external countermeasures tested so far have had limited success. Schaffner, Heysen, & Chetty, 2026


The term also needs precision. “Dark pattern” is a research, design, policy, and regulatory term, not a psychiatric diagnosis. It does not mean that every persuasive feature is deceptive, that every convenience-oriented default is harmful, or that every design optimized for engagement is a dark pattern. The boundary depends on what the interface does to the user’s ability to understand, choose, refuse, reverse, or leave.


What Are Dark Patterns?


The phrase “dark patterns” became a practical label for design choices that redirect user behavior toward outcomes favored by a service while making the user’s own preferences harder to express. Human-computer interaction research later developed more systematic taxonomies. Gray and colleagues analyzed practitioner-identified examples and organized recurring problems into families such as nagging, obstruction, sneaking, interface interference, and forced action. Gray et al., 2018


There is no single universally accepted definition. Different fields emphasize different elements: deception, manipulation, coercion, interference with autonomy, distortion of choice, consumer harm, or conflict between user and provider interests. A 2025 systematic literature review of 116 academic and regulatory sources found substantial definitional variation and proposed a broad legal formulation centered on a digital choice environment that is capable of distorting user behavior. Importantly, the review argued that designer intent and exploitation of a specific cognitive bias should not be treated as necessary elements of the definition. systematic review of dark-pattern definitions


Regulators increasingly use related language. The European Data Protection Board uses “deceptive design patterns” for interfaces and user journeys that can influence people toward unintended, unwilling, or potentially harmful decisions about personal-data processing. EDPB Guidelines 03/2022 The OECD uses “dark commercial patterns” for digital practices that steer, deceive, coerce, or manipulate consumers into decisions that may not be in their best interests. OECD, 2022


For psychology, the most useful working definition is functional: a dark pattern is a digital choice architecture that materially tilts behavior through deception, hidden information, asymmetrical friction, coercive pressure, or other design that weakens a person’s practical ability to make a free and informed choice.


Dark Patterns Are Part of Choice Architecture


Every interface has choice architecture. A button has a position. An option has a label. A setting has a default. A checkout has an order. A cancellation process has a number of steps. A notification arrives at a particular time. Even a deliberately neutral-looking interface still organizes information in some way.


That is why “choice architecture” and “dark pattern” are not synonyms. The UK Competition and Markets Authority describes online choice architecture as the digital environment in which people act and make decisions, including the presentation, placement, and design of options. Its evidence review explicitly notes that online choice architecture can help users by simplifying information, making journeys smoother, and surfacing relevant options. The same design tools can also be used in ways that create consumer or competition harms. CMA evidence review, 2022


This distinction matters because a psychologically effective interface is not automatically manipulative. A medication reminder can make an important action more salient. A security warning can intentionally interrupt a user before a risky action. A streaming service can remember where someone stopped. An accessibility setting can reduce friction. A clearly labeled default can simplify a complex configuration. These are examples of design influencing behavior without automatically becoming dark patterns.


The critical question is how the influence is structured. Does the design help a person carry out an understood goal, or does it make a provider-favored outcome disproportionately easy while hiding, confusing, or burdening the alternative? Does it support informed choice, or does it strategically exploit the fact that people have limited time, attention, and willingness to navigate friction?


Dark Patterns vs Persuasive Design


Persuasive design is broader than dark patterns. A product can encourage repeated use, completion, exercise, learning, saving, reading, or participation without deceiving the user. The target behavior may even be explicitly chosen by the user.


Dark patterns occupy a narrower territory in which influence is combined with deception, obstruction, asymmetry, concealed consequences, coercive pressure, or another feature that undermines the user’s ability to choose according to their own preferences. A “Continue” button that is visually prominent is not inherently a dark pattern. A “Continue” button may become part of a deceptive pattern when the prominent path commits the user to a materially different outcome than they reasonably understand, while the alternative is obscured or burdened.


This is also why not every engagement feature belongs in the dark-pattern category. Infinite scroll and autoplay can remove natural stopping cues and extend sequences of content, but their psychological mechanics are better analyzed as continuity and stopping-cue design. Classification as a dark pattern depends on context, disclosure, user control, and what the design is doing in the specific choice environment. See Infinite Scroll and Autoplay: How Endless Feeds Change Stopping Cues and Time Awareness.


The broader market incentive to attract and retain attention is covered separately in Attention Economy: How Digital Platforms Compete for Your Time and Focus. Attention competition can exist without deception, and deception can occur in a transaction that takes only seconds. The concepts overlap without collapsing into one another.


Dark Patterns vs Bad Usability


Poor usability can frustrate users without being a dark pattern. A confusing settings page may result from weak information architecture, technical debt, inconsistent design, or simple error. A dark-pattern analysis asks whether the confusion or friction systematically pushes behavior toward an outcome that benefits one side of the interaction at the expense of the user’s expressed preference or informed choice.


Intent can be difficult to establish from an interface alone. A user usually cannot know whether a designer deliberately made cancellation difficult or whether a product accumulated an awkward workflow over time. That is one reason newer legal and research definitions often focus on the structure and effects of the choice environment rather than requiring proof of a designer’s internal motive. systematic review of definitions, 2025


The practical consequence is useful: identify the concrete design mechanism first. Describe what is hidden, emphasized, preselected, delayed, repeated, or made harder. Then assess evidence about its behavioral effects and, where relevant, its legal status. Avoid turning every irritating interface into a claim about deliberate manipulation.


The Main Families of Dark Patterns


Taxonomies differ, and the same interface can contain more than one pattern. The five families developed by Gray and colleagues remain useful because they describe recurring ways that interfaces interfere with choice rather than relying on a long list of branded labels. Gray et al., 2018


Interface Interference


Interface interference changes the presentation of choices so that one option receives disproportionate visual or interactional advantage. A bright, large “Accept all” button paired with a low-contrast “Manage settings” link is a familiar example. The information may technically exist, but the interface allocates attention unevenly.


The mechanism is not that color “controls the brain.” Salient, spatially prominent, high-contrast, or immediately available options are easier to notice and act on. The alternative requires additional search, interpretation, or effort. In a repeated environment, small differences in friction and salience can accumulate into large differences in behavior.


Interface interference can also appear through confusing button labels, visual hierarchy that makes refusal look unavailable, or layouts that position a provider-favored action where users expect a neutral navigation control. Whether a specific implementation is deceptive depends on the full interaction, not on a single color or button in isolation.


Obstruction


Obstruction creates unnecessary difficulty when a person tries to exercise a choice that conflicts with the provider’s preferred outcome. The classic example is asymmetry between joining and leaving: subscription may require one or two clicks, while cancellation requires searching menus, navigating multiple screens, contacting support, or repeatedly rejecting retention offers.


The psychological effect of obstruction comes from effort. Every additional screen, wait, search step, form field, or ambiguous instruction creates another opportunity for a person to postpone, abandon, or forget the task. This is sometimes described as friction or “sludge.” The important feature is asymmetry: the burden is concentrated on the path the provider would rather the user not take.


The EU Digital Services Act specifically identifies making termination of a service more difficult than subscribing as an example relevant to its prohibition on interface design that deceives, manipulates, or materially distorts free and informed decisions. Digital Services Act, Article 25


Sneaking


Sneaking means material information, products, costs, permissions, or consequences enter the interaction in a way that is easy to miss. Examples include an add-on appearing in a basket, a recurring charge disclosed late, or important limitations appearing only after the user has invested substantial time in a transaction.


Timing is psychologically important. Information that appears after a person has searched, compared, entered details, and mentally committed to completion is encountered in a different decision context from information shown clearly at the beginning. Late disclosure can make the cost of backing out feel higher even when the monetary terms have not changed.


A large-scale study of approximately 11,000 shopping websites found 1,818 instances of 15 types of dark patterns across 1,254 sites, about 11.1% of the sample. The authors emphasized that their automated approach captured only part of the design space, particularly text-based patterns, so the estimate should not be treated as a universal prevalence rate for all websites or apps. Mathur et al., 2019


Forced Action


Forced action requires a user to perform an additional action to access or complete something they want. That can include creating an account, sharing information, granting a permission, subscribing to messages, or connecting another service.


A required step is not automatically manipulative. Some actions are genuinely necessary for security, payment, legal compliance, synchronization, or the core function the user requested. The dark-pattern question is whether the extra action is necessary, clearly explained, proportionate, and meaningfully optional where it should be optional.


The same logic applies to privacy. Asking for information can be legitimate. Conditioning an unrelated service on unnecessary data disclosure, or designing consent so that refusal is materially harder to express, raises a different issue.


Nagging and Repeated Requests


Nagging repeatedly interrupts a person with the same request or provider-preferred choice after the person has already declined, dismissed, or postponed it. Repetition can be effective because attention is finite and because declining itself has a cost. A person may eventually agree simply to stop the interruption.


The EU Digital Services Act explicitly identifies repeatedly requesting a choice that a user has already made, especially through interfering pop-ups, as a practice relevant to Article 25. Digital Services Act, Article 25


Repeated prompts also illustrate why users should not be described as passive victims of algorithms. A person can resist, ignore, change settings, leave, or adapt. At the same time, the interface controls how often resistance is required. The behavioral outcome emerges from the interaction between design, goals, context, habits, incentives, and individual circumstances.


Common Named Patterns in Real Interfaces


Popular names such as “confirmshaming,” “roach motel,” “hidden costs,” “preselection,” “trick questions,” “disguised ads,” “false hierarchy,” “scarcity,” and “urgency” are useful shorthand, but taxonomies vary. The same label can cover different mechanisms, and the same interface can fit several labels.


“Confirmshaming” uses emotionally loaded refusal language—for example, making the decline option imply irresponsibility, ignorance, or lack of care. The mechanism is social or self-evaluative pressure rather than a literal loss of choice.


A “roach motel” describes a process that is easy to enter and hard to leave. Subscription cancellation is the obvious example, but the principle can also apply to account deletion, marketing preferences, or withdrawal of consent.


“Trick questions” use wording, polarity, or button labels that make the intended meaning difficult to parse. A user may need to determine whether “No, do not cancel” means keep the service, stop cancellation, or decline a prompt. These interfaces increase the probability of an action error by making the mapping between intention and control unnecessarily ambiguous.


“Disguised ads” make commercial content resemble navigation, system messages, download buttons, or ordinary editorial content. Here the core problem is source recognition: the user may act before identifying that the element is advertising.


Scarcity and urgency require extra care. A truthful statement that two seats remain or a sale ends at midnight can be useful information. A fabricated countdown timer, false stock warning, or misleading claim that many other people are currently purchasing the same item is different. The dark-pattern status depends on truthfulness, context, and how the information is used.


How Dark Patterns Shape Attention


Attention is central to dark patterns because interfaces decide what competes for perception at the moment of choice. A screen can make one option dominant through position, size, contrast, animation, repetition, sequencing, or timing. It can also hide relevant information below a fold, behind an extra link, in dense text, or on a later screen.


This does not mean attention is a single resource that an interface “drains.” Selective attention, sustained attention, executive control, working memory, and task switching are related but distinct processes. Dark patterns most directly concern selective access to decision-relevant information: what the user notices, what they miss, and how much cognitive work is required to compare alternatives.


Attention can also be captured through interruption. A repeated modal dialog or prompt temporarily replaces the user’s task with the platform’s request. If the refusal path is cognitively harder than acceptance, the design combines attentional capture with decision friction.


The useful question is therefore not “Did the interface destroy attention span?” It is “How did the interface allocate salience and effort at this decision point?” That framing keeps the claim measurable and prevents momentary effects from being turned into unsupported claims about permanent neurological change.


For the broader problem of digital environments competing with ongoing goals, see Digital Distraction: How Phones, Apps, and Online Environments Compete for Attention.


Salience and Visual Asymmetry


Human attention is selective. People cannot inspect every element of a complex screen with equal depth. Designers therefore use hierarchy to help users understand what is primary, secondary, actionable, or informational.


A dark pattern can repurpose this normal design principle by making the provider-favored action salient while degrading the visibility of an alternative. The choice may remain formally available yet become behaviorally less accessible. This is especially important in fast, repetitive interactions where users reasonably rely on conventional visual cues rather than reading every line as if reviewing a legal contract.


The distinction between visibility and availability is crucial. An option can exist in the interface while still being hard to find. Regulatory and HCI discussions increasingly focus on whether choices are presented in a way that preserves meaningful control, not merely whether an alternative is technically somewhere on the screen.


Defaults, Preselection, and Inertia


Defaults matter because many decisions involve an implicit option: what happens if the user does nothing or accepts the preconfigured state. Defaults can reduce effort and can be beneficial. They can also shift outcomes when users interpret the default as recommended, fail to notice it, or decide that changing it is not worth the additional effort.


A prechecked marketing box, default data-sharing setting, automatically renewed plan, or preselected add-on can therefore shape behavior without any dramatic emotional manipulation. The effect may arise from inattention, status-quo preference, perceived endorsement, or simple friction.


This is also why a “default” is not automatically a dark pattern. The ethical and legal assessment depends on the domain, whether the default is disclosed, what consequences follow, whether active consent is required, and how easily the user can change the setting.


Friction, Effort, and Asymmetry


One of the strongest recurring mechanisms in online choice architecture is asymmetrical friction. When one action takes one click and the alternative takes six, the interface assigns different costs to the two preferences.


Friction is not always bad. High-stakes actions sometimes benefit from confirmation. Deleting an account, transferring money, changing a security credential, or publishing sensitive information may reasonably require additional steps. The psychological issue is whether the friction is aligned with user protection or strategically placed to block an unwanted business outcome.


A 2026 economic experiment introduced a “transaction test” for deceptive online interfaces and found that tested dark patterns increased choice inconsistency by roughly 6–12 percentage points, depending on the condition, compared with a plain-design control. A transparency-oriented remedy in that study did not eliminate the effect. The transaction test, 2026


Wording, Framing, and Emotional Pressure


Interfaces do not only organize buttons; they organize meaning. A refusal labeled “No thanks” is psychologically different from one labeled “No, I prefer to stay unprotected.” Both may lead to the same technical state, but the second adds a judgment about the user’s identity or values.


This kind of framing can introduce guilt, fear, embarrassment, anticipated regret, or self-presentation concerns. It can also make the user spend additional effort translating persuasive copy back into the actual underlying choice.


The evidence does not justify assuming that every person reacts the same way. Some users recognize the tactic immediately and resist it. Others may be indifferent. Context matters. The scientifically stronger claim is probabilistic: framing can change the distribution of choices across groups, and particular dark-pattern experiments have demonstrated sizable shifts under controlled conditions.


Scarcity, Urgency, and Social Information


Scarcity and urgency are psychologically potent because they can alter the perceived cost of delay. A message that an offer expires soon, inventory is low, or many people are viewing an item can make postponement feel risky.


These cues can be legitimate. Real inventory is useful. A real deadline is useful. Genuine social information can help people assess popularity. The design problem emerges when the claim is false, unverifiable, strategically incomplete, or presented with enough pressure that users cannot reasonably evaluate the decision.


Mathur and colleagues identified urgency and social-proof patterns in online shopping and also documented deceptive examples, including messages whose underlying claims did not correspond to observable reality. Mathur et al., 2019 The FTC likewise describes false countdowns, misleading scarcity, and similar practices in its staff report on dark patterns. FTC, 2022


Information Overload and Strategic Complexity


A platform can influence choice by adding information as well as hiding it. Long text, multiple settings, nested menus, repeated explanations, and dense option sets can make a decision so costly that the user chooses the fastest available path.


Information overload should not be treated as proof of manipulation. Complex products genuinely require complex information. Privacy and financial choices can involve many dimensions. The question is whether complexity reflects the underlying decision or whether the interface selectively simplifies the provider-favored option while making the alternative difficult to understand.


The CMA’s evidence review highlights how complex language, choice overload, ranking, defaults, scarcity cues, and other forms of online choice architecture can influence consumer behavior. It also emphasizes that many of these practices can be beneficial in some contexts and harmful in others. CMA evidence review, 2022


What Experimental Research Shows


The strongest evidence that dark patterns can change behavior comes from randomized and controlled experiments, not from the mere observation that a manipulative-looking interface exists.


Luguri and Strahilevitz conducted experiments with U.S. participants using a fictitious identity-protection offer. In one study, 11.3% of participants in the control condition accepted the offer, compared with 25.8% in a mild dark-pattern condition and 41.9% in an aggressive condition. The aggressive condition also produced more anger, more dropout, and a lower reported sense of freedom to refuse. Luguri & Strahilevitz, 2021


Those results are important, but they should not be treated as a universal effect size. The study used a particular offer, interface sequence, participant sample, and experimental context. Other patterns produce smaller effects, and some designs fail to move behavior reliably.


A 2026 systematic review of user experiments synthesized this broader literature and found substantial agreement that deceptive/manipulative patterns alter behavior, with large variation in effect magnitude. It also found that correlations between pattern effects and personal characteristics were not consistently strong across studies and that many tested countermeasures did not reliably neutralize the effects. Schaffner, Heysen, & Chetty, 2026


This supports a useful conclusion: dark patterns can matter even when users are intelligent, experienced, and aware that interfaces are designed. Susceptibility is not evidence of gullibility. Digital decisions are made under limited time and attention, and even small asymmetries can influence aggregate behavior when millions of interactions occur.


Cookie Banners and Privacy Choices


Cookie and privacy-consent interfaces provide especially clear experimental evidence because researchers can vary the design while measuring consent choices.


Nouwens and colleagues studied consent management platforms on popular UK websites and then tested common banner designs experimentally. In their field experiment, removing the opt-out button from the first page increased consent by about 22–23 percentage points, while placing more granular controls on the first page reduced consent by about 8–20 percentage points. Nouwens et al., 2020


These results do not show that users have no privacy preferences. They show that the path through which a preference must be expressed changes the probability that it is expressed. A person who would decline tracking under a symmetric interface may accept under an interface that requires extra navigation, just as another person may accept quickly because that genuinely matches their preference.


The distinction between preference and observed click is central to dark-pattern research. Interface behavior is an outcome produced jointly by what the person wants and what the environment makes easy, salient, comprehensible, or costly.


Prevalence: How Common Are Dark Patterns?


There is no single reliable percentage for “how much of the internet uses dark patterns.” Studies examine different domains, platforms, countries, time periods, and definitions. Detection methods also differ. A crawler that can read text will miss visual hierarchy, animation, timing, and interaction sequences.


The 2019 Mathur study remains a foundational prevalence analysis in e-commerce: 1,818 identified instances appeared on 1,254 of roughly 11,000 shopping websites. The researchers explicitly described this as a lower-bound estimate within their method. Mathur et al., 2019


Research with children’s apps demonstrates that manipulative design also occurs outside conventional web shopping. Radesky and colleagues coded apps used by 160 children ages 3–5 and found manipulative design features such as fabricated time pressure, navigation constraints, parasocial pressure, advertising pressure, and purchase pressure. Only 20% of coded apps had no manipulative design features. The study was cross-sectional and used a convenience sample, so it establishes prevalence in that sample rather than a universal rate or causal mental-health effect. Radesky et al., 2022


A 2025 systematic review of legal and HCI scholarship concluded that the field still struggles with definition, measurement, and enforcement boundaries. That matters for prevalence claims: a wider definition will classify more interfaces, while a narrow deception-based definition will classify fewer. Yi & Li, 2025


Who Is Most Vulnerable?


It is tempting to divide users into “susceptible” and “resistant” groups, but current evidence supports a more contextual view.


Zac and colleagues ran a large online experiment with 2,500 UK participants and found strong evidence that people across demographic groups were susceptible to dark patterns. Common proxies for vulnerability, including income, education, and age, showed only weak evidence of materially changing susceptibility in their design. They also found that adding a required payment action reduced the effectiveness of the patterns, highlighting how transaction structure can matter as much as demographic category. Zac et al., 2025


That finding does not mean vulnerability is irrelevant. Children, people with particular cognitive or accessibility needs, people under time pressure, people in financial distress, and users with limited domain knowledge may face different risks depending on the interaction. It means researchers should not assume vulnerability from a demographic label without evidence.


The Radesky study also found that children from lower-socioeconomic-status households in its sample encountered apps with more manipulative design features. That is an exposure inequality, not proof that those children were psychologically more persuadable. Radesky et al., 2022


Dark Patterns and Mental Health


Dark patterns can produce annoyance, frustration, perceived coercion, regret, privacy loss, unwanted purchases, or difficulty disengaging from a service. Experimental evidence supports some of these immediate psychological and behavioral outcomes.


Evidence for direct clinical mental-health consequences is much thinner. A user feeling angry after resisting an aggressive interface is not the same as having an anxiety disorder or depression. A manipulative app design is not itself a psychiatric exposure with an established diagnostic pathway. Research discussing mental-health implications often identifies plausible risks and vulnerable contexts rather than demonstrating that dark patterns cause a clinical disorder.


This distinction is essential for YMYL accuracy. Dark-pattern research supports claims about decision behavior, autonomy, attention, effort, annoyance, trust, privacy, spending, and related outcomes more strongly than it supports claims about psychiatric diagnosis.


If digital use is associated with persistent distress or functional impairment, the relevant question is the person’s broader pattern of symptoms, context, and functioning. It should not be reduced to the presence of a manipulative interface.


Do Dark Patterns Cause “Addiction”?


Dark patterns can encourage continued engagement or make stopping harder, but that does not establish a behavioral addiction.


Some interfaces remove stopping cues, repeatedly prompt re-entry, make cancellation difficult, or reward repeated checking. Those mechanisms can contribute to repeated behavior. The clinical word “addiction,” however, requires a separate diagnostic and evidence framework. A design feature cannot by itself diagnose an individual, and a measured increase in clicks or time does not prove a disorder.


This distinction also protects the science from viral explanations such as “dopamine addiction” or “brain hijacking.” Reward learning, habits, salience, social motivation, and reinforcement can all be relevant without turning dopamine into a synonym for desire or digital harm.


Dark Patterns and the Attention Economy


The attention economy describes incentives that arise when attention is scarce and strategically valuable. Dark patterns describe a subset of interface practices that distort choice through deceptive, obstructive, coercive, or asymmetrical design.


The two concepts intersect when a platform uses a manipulative interface to extend engagement, obtain notifications permission, secure data consent, prevent cancellation, or make a return visit more likely. But attention economy is the broader incentive structure; dark pattern is the interface-level mechanism.


This separation matters for causal claims. A platform may compete intensely for attention using recommendation quality, entertainment, useful notifications, or creator content without employing a dark pattern in a particular interaction. Conversely, an e-commerce checkout can use a dark pattern to add a paid service even when the entire interaction lasts less than a minute.


See Attention Economy: How Digital Platforms Compete for Your Time and Focus for the broader economic and psychological framework.


Are Infinite Scroll and Autoplay Dark Patterns?


Not automatically.


Infinite scroll removes a page boundary and can reduce stopping cues. Autoplay begins the next item without requiring a new initiation. Both can make continuation easier. That is a design effect that can be studied directly.


Calling either feature a dark pattern requires an additional judgment about how it functions in context. Is the user clearly informed? Can autoplay be disabled? Is the feature part of a deceptive or coercive choice? Does it materially impair an expressed preference to stop, refuse, or control the sequence? Different researchers and regulators may draw the boundary differently.


For the evidence on stopping cues, time awareness, and continuous content specifically, see Infinite Scroll and Autoplay: How Endless Feeds Change Stopping Cues and Time Awareness.


Why Intent Is a Difficult Criterion


The original language around dark patterns often implied deliberate deception. In many real cases, deliberate strategy may indeed exist. But intent is difficult to infer from a screen.


A problematic interface can emerge from A/B testing, business metrics, legacy code, fragmented teams, local optimization, copied industry conventions, or explicit attempts to increase conversion. The user usually sees the result, not the organizational history that produced it.


That creates a measurement problem. If dark-pattern classification requires proof that a specific designer intended to manipulate, many harmful interfaces become impossible to classify without internal evidence. If classification ignores intention entirely, researchers risk labeling any badly designed interface a dark pattern.


Recent systematic legal work proposes resolving this by focusing first on whether the digital choice environment is capable of distorting user behavior, leaving intent as a separate evidentiary question when a legal rule or ethical analysis requires it. systematic review, 2025


For readers, the practical rule is simpler: describe observable design before making claims about motive.


Regulation: Are Dark Patterns Illegal?


There is no global rule under which every interface called a dark pattern is automatically illegal. Legal status depends on jurisdiction, sector, conduct, user group, data practice, and the specific rule being applied.


In the European Union, Article 25 of the Digital Services Act prohibits providers of online platforms from designing, organizing, or operating online interfaces in ways that deceive or manipulate recipients or otherwise materially distort or impair their ability to make free and informed decisions, subject to the provision’s relationship with other EU rules. The article specifically mentions giving greater prominence to certain choices, repeatedly requesting a choice already made, and making service termination more difficult than subscription. Regulation (EU) 2022/2065, Article 25


For personal data, the European Data Protection Board’s final Guidelines 03/2022 describe deceptive design patterns in social-media interfaces and connect them with GDPR principles such as fairness, transparency, accountability, and data protection by design and by default. EDPB Guidelines 03/2022


In the United States, the Federal Trade Commission has described dark patterns that can trick or manipulate consumers into purchases, subscriptions, data sharing, or privacy choices and has connected such practices with existing consumer-protection law and enforcement. FTC, Bringing Dark Patterns to Light


The OECD’s policy analysis surveys consumer harms and regulatory approaches across jurisdictions rather than establishing a single binding global standard. OECD, 2022


The practical conclusion is jurisdiction-specific: “dark pattern” is a useful descriptive category, while legality must be assessed under the rules that govern the actual conduct.


Privacy Dark Patterns


Privacy interfaces are a major research area because the design must translate an abstract preference—how much data a person wants to share—into concrete clicks.


A privacy screen can influence behavior through preselection, unequal button prominence, additional steps to refuse, confusing toggles, repeated prompts, bundled purposes, or information presented only after the user has already committed to an action. The EDPB’s guidelines organize deceptive design patterns in this domain around recurring problems such as overloading, skipping, stirring, hindering, fickle interfaces, and information left out of context. EDPB Guidelines 03/2022


The Nouwens experiment shows why this is psychologically important: changing the availability and granularity of controls changed consent behavior substantially. Nouwens et al., 2020


A privacy decision therefore cannot be interpreted solely from the final click. Consent behavior reflects both preference and architecture.


Subscription and Cancellation Dark Patterns


Subscription interfaces combine several powerful design dimensions: defaults, recurring payments, effort asymmetry, reminders, timing, and the difference between immediate and delayed consequences.


An easy signup followed by difficult cancellation creates obstruction. A trial that converts to payment without clear, timely disclosure can involve sneaking or inadequate disclosure. Repeated retention screens can become nagging. A large “Keep my plan” button paired with a buried cancellation link can create interface interference.


These mechanisms can be studied without assuming that every subscription is exploitative or that every retention offer is manipulative. A service can legitimately explain what a user will lose, offer a discount, or ask for feedback. The boundary is whether the user can still understand and execute the choice without deceptive or disproportionate burden.


Shopping and Financial Consequences


Dark patterns can influence purchases through hidden costs, preselected extras, misleading scarcity, countdown timers, social-proof messages, or difficult reversal. The foundational Mathur crawl focused on shopping websites for precisely this reason. Mathur et al., 2019


When the primary question is interface behavior—how the design steers a click, consent, subscription, or choice—the topic belongs to digital choice architecture. When the primary question is spending, debt, pricing psychology, or financial decision-making, the stronger conceptual home is economic psychology.


This boundary prevents a common SEO problem: treating “dark patterns” as a catch-all article for every downstream consequence. The interface mechanism and the financial outcome are related, but they are not the same search intent.


Children and Manipulative Design


Children deserve separate attention because developmental stage changes reading ability, impulse control, commercial understanding, social inference, and the ability to identify persuasive intent.


Radesky and colleagues documented manipulative design features in apps used by preschool-aged children, including characters applying social pressure, fabricated urgency, lures that prolonged play, advertisements, and purchase pressure. Radesky et al., 2022


The study does not establish that these features cause a particular psychiatric outcome. It does show that manipulative design is present in environments used by very young children and that developmental context matters when evaluating what counts as meaningful choice.


A child-centered analysis should therefore ask whether the user can understand the commercial or data consequence, distinguish advertising from content, recognize a false or emotional claim, and reverse the action. The same interface can pose different practical risks across developmental stages.


Why Awareness Alone Is Not a Complete Solution


A common response to dark patterns is “teach users to recognize them.” Recognition can help, but it transfers the burden of correcting design onto the person encountering it.


The 2026 systematic review of user experiments found that interventions intended to mitigate dark-pattern effects have often produced limited or inconsistent protection. Schaffner, Heysen, & Chetty, 2026 The 2026 transaction-test experiment likewise found that a transparency-oriented remedy did not eliminate the measured inconsistency effects in its tested conditions. The transaction test, 2026


That does not mean user education is useless. It means structural design matters. A person should not need advanced HCI knowledge to cancel a subscription, refuse unnecessary data sharing, or identify the actual total price of a transaction.


The strongest protection combines user awareness with fair defaults, symmetrical choices, truthful information, easy reversal, regulatory standards, and product-level design that does not require constant vigilance.


How to Recognize a Dark Pattern


A useful recognition method is to inspect the relationship between alternatives rather than focusing on one suspicious button.


First, compare visibility. Are accept and decline options similarly easy to find, or is one visually dominant while the other is hidden in text, color, or hierarchy?


Next, compare effort. How many steps does each path require? Is starting easy while stopping, canceling, refusing, or deleting requires substantially more work?


Then, inspect timing. Was a material cost, recurring charge, data use, or limitation disclosed before commitment, or only after the user had already invested time and information?


Check the default. What happens if the user does nothing? Is an optional choice already selected? Is a recurring state enabled without a clearly understood action?


Examine wording. Does the interface describe the actual consequence, or does it use guilt, fear, double negatives, ambiguous buttons, or emotionally loaded refusal language?


Look at repetition. If the user has already chosen, does the system repeatedly demand that the choice be made again?


Finally, test reversibility. Can the person undo the action through a path comparable to the one used to take it?


No single item proves a dark pattern. Together, these questions reveal whether the architecture supports or distorts the user’s ability to express a preference.


How Users Can Reduce the Influence of Dark Patterns


The most useful strategies increase decision time and reduce asymmetric friction at the moments when the stakes matter.


For purchases and subscriptions, inspect the final total, recurring terms, preselected extras, trial conversion date, and cancellation route before completing the transaction. For privacy choices, use the detailed settings when they materially affect your preference rather than treating the largest button as the only option. For urgency claims, verify whether the deadline or scarcity is real when the decision matters.


When a choice is consequential, temporarily leaving the interface can be useful. Compare the offer in another source, read the policy or price outside the pop-up, or return after the immediate urgency has passed. This reduces the power of timing and commitment within a single flow.


Screenshots can also be practical when a price, timer, cancellation route, or disclosure changes. They create a record of what the interface presented at the moment of choice.


System-level tools—notification controls, browser privacy settings, password managers, app-permission managers, payment alerts, and platform subscription dashboards—can reduce repeated exposure to some manipulative flows. They are supports, not guarantees.


The goal is not to treat every interface as hostile. It is to preserve enough time and clarity to notice when the design and your own goal are moving in different directions.


What Ethical Interface Design Looks Like


Ethical interface design makes the consequences of a choice legible and keeps meaningful alternatives practically available.


That usually means truthful claims, timely disclosure, understandable labels, proportional friction, reversible actions, clear defaults, and comparable prominence for materially opposing choices. It means avoiding a design in which the user must decode business persuasion before they can express a basic preference.


The EDPB recommends privacy interfaces that support fairness, transparency, accountability, and data protection by design and by default. EDPB Guidelines 03/2022 The FTC’s report similarly highlights design practices that obscure terms, make cancellation difficult, disguise advertising, or steer privacy decisions. FTC, 2022


Ethical design does not require removing all persuasion. Products have goals, and users often want guidance. The standard is that guidance should remain compatible with understanding, choice, and reversal.


How Researchers Test Dark Patterns


Dark-pattern research uses several methods, each answering a different question.


Large-scale audits and crawls estimate prevalence, but they can miss visual and interactive elements. Laboratory and online experiments can isolate causal effects of a design change, but the setting may differ from everyday use. Field experiments increase realism but are harder to control. Qualitative studies can reveal how users interpret pressure or deception but do not estimate population effect sizes by themselves.


Systematic reviews are therefore especially valuable because they show which findings replicate across methods and where the literature remains fragmented. The 2026 CHI review focuses specifically on user experiments and behavioral effects. Schaffner, Heysen, & Chetty, 2026 The 2025 Computer Law & Security Review synthesis maps definitions, harms, regulatory paradigms, and proposed solutions across law and HCI. Yi & Li, 2025


The evidence base is now substantial enough to reject the idea that dark patterns are merely an internet buzzword. It is still heterogeneous enough that claims should identify the specific pattern, context, population, and outcome.


What the Evidence Does Not Show


Dark-pattern research does not show that every user is manipulated every time. It does not establish one universal effect size. It does not prove that every high-engagement interface is deceptive. It does not justify inferring a company’s motive from a screenshot alone. It does not show that brief exposure permanently changes the brain or “destroys attention.”


It also does not show that awareness completely protects users. Experimental work demonstrates that people can remain influenced even when interfaces look obviously strategic, and intervention effects vary.


Finally, it does not justify collapsing consumer behavior, attention, privacy, spending, and clinical mental health into a single outcome. These are different constructs and require different measures.


Precision strengthens the case against genuinely deceptive design because it separates documented mechanisms from viral exaggeration.


Frequently Asked Questions


What is a dark pattern in psychology?


A dark pattern is a digital choice architecture that changes the probability of a user’s decision through deceptive, obstructive, coercive, hidden, or asymmetrical design. Psychology helps explain how salience, defaults, framing, effort, repetition, time pressure, and context interact with user goals and attention. “Dark pattern” is not a psychiatric diagnosis.


Are dark patterns the same as persuasive design?


No. Persuasive design includes any interface intentionally created to encourage behavior. It can support goals that users explicitly want. Dark patterns are the subset in which influence is paired with deception, hidden consequences, obstruction, coercive pressure, or another distortion of meaningful choice.


Are all defaults dark patterns?


No. Defaults can reduce complexity and support users. Their status depends on context, disclosure, consequences, and whether the default is appropriate for the decision. Some legal domains require active consent rather than preselection.


Is a big “Accept” button automatically a dark pattern?


No. Visual prominence is one piece of the interaction. It becomes more concerning when the opposing choice is materially harder to find or execute, the consequence is poorly disclosed, or the hierarchy systematically steers users away from an informed preference.


Is infinite scroll a dark pattern?


Not automatically. Infinite scroll removes natural stopping boundaries and can make continuation easier. Whether a particular implementation counts as a dark pattern depends on the surrounding choice architecture, disclosure, control, and whether it deceives or materially impairs an expressed choice. See Infinite Scroll and Autoplay.


Do dark patterns work because of dopamine?


There is no single “dopamine dark-pattern mechanism” that explains the evidence. Dark-pattern effects can involve attention, learned expectations, salience, effort, framing, emotion, default effects, and social information. Dopamine is involved in many normal learning and motivational processes and should not be used as a one-word explanation for digital persuasion.


Can dark patterns cause ADHD?


Current evidence does not establish that dark patterns cause ADHD. ADHD is a neurodevelopmental disorder with a separate diagnostic framework. Research showing that an interface changes attention or choice in the moment is not evidence that it causes ADHD.


Can dark patterns cause anxiety or depression?


Research documents behavioral effects, annoyance, perceived coercion, privacy and financial harms, and related experiences. Direct causal evidence that dark patterns themselves cause a clinical anxiety or depressive disorder is limited. Clinical symptoms and diagnoses require separate evidence.


Are dark patterns illegal in the United States?


Some practices described as dark patterns can violate U.S. consumer-protection, privacy, subscription, advertising, or sector-specific requirements, but “dark pattern” is not a single nationwide legal switch that makes every disputed interface automatically unlawful. The FTC evaluates conduct under applicable law and has published detailed guidance and enforcement-oriented analysis. FTC, 2022


Are dark patterns illegal in the European Union?


EU law contains explicit restrictions relevant to deceptive interface design. Article 25 of the Digital Services Act prohibits covered online platforms from using interface design that deceives, manipulates, or materially distorts or impairs free and informed decisions, subject to the provision’s scope and interaction with other EU law. GDPR-related deceptive design is also addressed by the EDPB. Digital Services Act EDPB Guidelines


Why do people fall for dark patterns even when they know about them?


Knowing that influence exists does not remove the practical costs of attention, time, interpretation, and repeated resistance. Experiments show that interface asymmetries can still shift aggregate choices. The 2026 systematic review also found that many tested mitigation strategies have not reliably eliminated effects. Schaffner, Heysen, & Chetty, 2026


What is the best way to avoid dark patterns?


There is no universal technique. The most reliable practical approach is to slow consequential decisions, inspect defaults and totals, compare the effort required by opposing choices, verify urgency claims, review privacy and subscription settings, and use system-level controls where helpful. Structural protections and fair interface design matter alongside individual awareness.


Conclusion: Digital Choice Is Partly Designed


Digital decisions do not occur in empty space. Interfaces determine which alternatives appear first, which are salient, which are preselected, which require extra work, when consequences are disclosed, and whether a refusal must be repeated.


That does not make users powerless. People bring goals, experience, preferences, skepticism, habits, and social context to every interface. Dark-pattern research shows something more specific: the architecture of a choice can change behavior, sometimes substantially, even when the formal set of options remains the same.


The strongest evidence now comes from converging HCI experiments, large-scale audits, systematic reviews, and regulatory evidence. It supports careful claims about behavioral steering, consent, friction, attention, autonomy, and consumer harm. It does not support turning every digital inconvenience into manipulation or every behavioral effect into a clinical disorder.


The practical standard is clear. A fair digital interface should make it possible to understand the choice, express a preference, refuse without disproportionate burden, and reverse a decision through a reasonable path. When design systematically weakens those capacities in order to tilt behavior, the interface itself becomes part of the psychological mechanism.


Related Articles



References


Bogliacino, F., Pejsachowicz, L., Liva, G., & Lupiáñez-Villanueva, F. (2026). The transaction test: An experimental method for assessing online interfaces. Journal of Economic Behavior & Organization, 244, 107469.





Federal Trade Commission. (2022). Bringing Dark Patterns to Light.


Gray, C. M., Kou, Y., Battles, B., Hoggatt, J., & Toombs, A. L. (2018). The dark (patterns) side of UX design. Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems.



Luguri, J., & Strahilevitz, L. J. (2021). Shining a light on dark patterns. Journal of Legal Analysis, 13(1), 43–109.


Mathur, A., Acar, G., Friedman, M. J., Lucherini, E., Mayer, J., Chetty, M., & Narayanan, A. (2019). Dark patterns at scale: Findings from a crawl of 11K shopping websites. Proceedings of the ACM on Human-Computer Interaction, 3(CSCW), Article 81.


Nouwens, M., Liccardi, I., Veale, M., Karger, D., & Kagal, L. (2020). Dark patterns after the GDPR: Scraping consent pop-ups and demonstrating their influence. Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems.


OECD. (2022). Dark commercial patterns. OECD Digital Economy Papers, No. 336.


Radesky, J., Hiniker, A., McLaren, C., Akgun, E., Schaller, A., Weeks, H. M., Campbell, S., & Gearhardt, A. N. (2022). Prevalence and characteristics of manipulative design in mobile applications used by children. JAMA Network Open, 5(6), e2217641.


Schaffner, B., Heysen, L., & Chetty, M. (2026). A systematic review of user experiments on the effects of dark patterns. Proceedings of the 2026 CHI Conference on Human Factors in Computing Systems, 1–17.



Zac, A., Huang, Y.-C., von Moltke, A., Decker, C., & Ezrachi, A. (2025). Dark patterns and consumer vulnerability. Behavioural Public Policy.

 
 
bottom of page