The Impact of Dark Patterns on the Validity of Digital Consent and the Possibility of Declaring Digital Contracts Invalid: A Comparative Legal Analysis
PRAKKOM.AZ, Digital Law Lab və TechLaw.az tərəfindən təşkil olunan Rəqəmsal Hüquq üzrə Akademik Məqalə Müsabiqəsinin I yer qalibi Fidan Yusiflinin rəqəmsal mühitdə manipulyativ dizayn üsullarının (Dark Patterns) istifadəçi razılığına, rəqəmsal müqavilələrin etibarlılığına və onların etibarsız hesab edilməsi imkanlarına, eləcə də bu sahədə Azərbaycan və beynəlxalq hüquqi tənzimləməyə həsr olunmuş tədqiqat işini təqdim edirik.
Məqalə: THE IMPACT OF DARK PATTERNS ON THE VALIDITY OF DIGITAL CONSENT AND THE POSSIBILITY OF DECLARING DIGITAL CONTRACTS INVALID: A COMPARATIVE LEGAL ANALYSIS
Abstract
This article examines the legal significance of dark patterns in the formation of digital consent and the validity of digital contracts. It addresses the question of when manipulative interface design may constitute a legally relevant defect of will and when it should remain a regulatory or consumer-protection issue. The study employs a comparative legal analysis of the European Union, the United States, and the Republic of Azerbaijan, alongside an examination of fraud, mistake, information duties, and consent. It argues that the presence of a dark pattern should not, by itself, render a transaction invalid. Its legal significance should instead be determined by the mechanism of influence, the accessibility and presentation of material information and alternatives, the effect on the user’s decision, and the material and causal connection between the manipulation and the transaction. On this basis, the article proposes a five-element functional test and a three-level framework of legal consequences. It further identifies gaps in Azerbaijani legislation and proposes introducing a specific provision on manipulative digital design into the Law on Consumer Protection, while preserving existing civil-law grounds for invalidity and freedom of legitimate commercial design.
Key words: dark patterns, digital consent, validity of transactions, defects of consent, consumer protection, digital contracts.
Introduction
In the modern digital environment, user interface design is no longer merely a tool for ensuring ease of use. The architecture of a digital environment can influence the sequence of a user's actions and which options they ultimately choose. This is particularly evident when subscribing to services, giving consent to the processing of personal data, and entering into digital contracts.
This role of the interface has become particularly significant as digital technologies have become increasingly integrated into everyday life. Recent research indicates that contemporary social communication is virtually impossible without the use of the digital environment.The information obtained enables platforms to develop personalised forms of interaction and adapt the interface to the characteristics of user behaviour.
The influence of design on user behaviour is not, in itself, unlawful. Richard Thaler, developing the concept of nudge, argued that choice architecture can help individuals make more informed decisions. At the same time, he drew attention to sludge, namely artificial obstacles that make it more difficult for users to take an action they wish to take. This distinction is relevant to the legal assessment of digital design. If an interface merely facilitates choice, it does not necessarily interfere with user autonomy. If, by contrast, it deliberately creates obstacles to one option or directs the user towards another contrary to their presumed interests, the question arises as to the permissible limits of such influence.
It is in this context that the concept of dark patterns emerged. In the present study, dark patterns are understood as user interface design choices that, through manipulation, direction or deception, induce users to take unintended and potentially harmful actions.
Harry Brignull first proposed the idea of ethically questionable approaches in 2010, defining dark patterns as follows: ‘a user interface that has been carefully designed to trick users into doing certain things... they are not mistakes, they are carefully designed based on a deep understanding of human psychology and without regard for the user’s interests’.4 Initially, this concept was used primarily in the field of UX design and e-commerce, but it subsequently became the subject of research in the areas of consumer protection, personal data and digital regulation
For legal analysis, it is important to recognise that dark patterns may affect users in different ways. Some practices make refusal or cancellation more difficult, others conceal material information, while others create a false sense of urgency or visually emphasise one choice, making the alternative less noticeable. Therefore, assessing the legal significance of dark patterns requires understanding the specific mechanism of influence. Classifying such practices makes it possible to determine how a particular interface may affect the formation of a legally significant expression of will and the subsequent validity of consent.
The academic literature does not provide a single classification of dark patterns, due to the continuous development of digital technologies and new ways of influencing user behaviour. This study uses the classification proposed by the OECD in its 2022 report Dark Commercial Patterns, as it systematises the main forms of manipulative digital design and their impact on consumer choice. The main types of such practices are presented in Table 1.

The classification above illustrates the range of mechanisms through which user behaviour can be influenced. The key issue, however, is not simply how these practices can be categorised, but whether they are capable of materially affecting the formation of a user’s will.
The impact of interface manipulation on users’ expressions of will is also supported by empirical research in behavioural economics. For example, a study by Blake et al. conducted on the StubHub platform found that displaying mandatory fees only at the final stage of the purchasing process significantly altered consumer behaviour. Compared with users who were shown the fees in advance, consumers in the Back-end Fee group were 14.1% more likely to make at least one purchase during the ten-day period. Among those who made a purchase, the average value of the tickets purchased was also 5.42% higher. The authors attribute this effect to the reduced salience of the final price, which encouraged consumers to select more expensive tickets.
Interface design can thus influence the formation of a person's internal will before a legally significant act is performed, yet behavioural influence and a legally relevant defect of will are not the same thing. An interface that increases the likelihood of a particular action does not automatically amount to fraud or material mistake. It is precisely this distinction that the substantial body of research on dark patterns and consumer behaviour has largely left unaddressed, as its implications for the validity of consent in civil law remain considerably underexplored.
Accordingly, this article adopts a narrower approach. When does the architecture of digital choice become legally relevant, and which existing legal doctrines can address it? The comparative analysis of the EU and the US aims to identify criteria applicable within Azerbaijani law, not merely to describe different regulatory models. I argue that dark patterns should be assessed functionally. Not every manipulation is a legal defect, but a material and causally connected distortion of free or informed choice should carry legal significance. This departs from two common responses: treating dark patterns as inherently unlawful, or leaving them to soft guidance alone.
Theoretical Foundations of the Validity of Consent and the Invalidity of Transactions
In civil law doctrine, the invalidity of a contract is understood as the absence, in an act performed in the form of an agreement, of the characteristics required of a legal fact capable of producing the civil-law consequences intended by the parties. One circumstance capable of rendering a transaction invalid is a discrepancy between the parties’ will and its expression.
Depending on the nature of the defect, invalid transactions are classified as void and voidable. A void transaction is invalid by virtue of the fact that it was entered into, whereas a voidable transaction remains legally effective until it is challenged by an authorised person and declared invalid in accordance with the prescribed procedure. The concept of voidability is particularly important for the subsequent analysis, as its grounds include, among others, transactions entered into under the influence of mistake or fraud.
Will and its declaration are interrelated but distinct elements of a transaction. While will reflects a person’s internal decision to achieve a particular legal result, its expression constitutes the external manifestation of that decision. For a transaction to produce the intended legal consequences, the expressed will must correspond to the person’s actual decision. Factors that distort either the formation of will or its expression may therefore affect the validity of the transaction.
Mistake is one such factor. It renders a transaction challengeable only where it is material to the decision to enter into it, concerning, for instance, the nature of the transaction, the identity or material characteristics of its subject matter, or another material circumstance, while minor errors do not meet this threshold.
Fraud differs from mistake in that it involves intentionally misleading a person to induce them to enter into a transaction. It may concern the transaction’s terms, including the rights and obligations arising from it, the quantity or quality of its subject matter, or the type and manner of providing a service, as well as other circumstances capable of influencing the formation of the party’s will.
Fraud or mistake alone does not suffice for the legal assessment of a transaction; two further elements are required. The relevant circumstance must be causally linked to the conclusion of the transaction, and it must be shown to have materially affected the formation of the person's will and led them to perform the legally significant act. It is this combination of causation and materiality that distinguishes a legally relevant defect of will from an ordinary mistake or other circumstance that did not decisively shape the decision.
The link between the validity of a transaction and the influence exerted on the declaration of will has deep historical roots. Roman private law addressed situations in which a person’s decision to enter into a transaction was shaped by deceptive conduct. A central concept was dolus (fraud), while dolus in contrahendo (fraud in the conclusion of a contract) specifically concerned the formation of a transaction. Roman jurists distinguished between dolus malus (malicious fraud) and dolus bonus (permissible trickery or exaggeration falling short of legally relevant fraud). Ulpian, conveying the definitions of Servius and Labeo, cites Labeo’s formulation that dolus malus encompasses trickery, deception or stratagem intended to mislead another person. Not every form of influence was regarded as dolus malus. Legal significance attached only to influence that caused the transaction and without which the person would not have made the relevant decision.
A particularly illustrative example is provided by Cicero in De officiis (de off., 3,14,58).9 He recounts the case of Canius, a Roman who wished to purchase an estate in Sicily. The seller, Pythius, created a false impression of the estate’s attractiveness by arranging in advance for fishermen to display a large catch in the surrounding area. Having witnessed this, Canius decided to purchase the estate and entered into the transaction, only to discover that the area was virtually devoid of fish. Cicero associates the case with fraud and the formula of Aquilius Gallus, cum esset aliud simulatum, aliud actum (‘when one thing was represented, while another was done’). The example demonstrates that fraud may arise not from any defect in the external expression of will, but from circumstances preceding it that influenced the decision to enter into the transaction. In Canius’s case, Pythius’s false representation of the estate’s qualities directly caused the transaction to be concluded.
What determines the legal defect is not the presence of a dark pattern as such, but its material influence on the user's decision and its causal link to the conclusion of the transaction.
The European Approach to the Regulation of Dark Patterns
For the European approach to dark patterns, the key issue is not the use of a particular interface element itself, but its impact on the user’s ability to make an autonomous and informed decision. Legal assessment therefore depends on the nature of the influence, the availability of alternatives, and the circumstances in which the user gives consent.
The first level of regulation concerns the GDPR and its requirements for consent to the processing of personal data. Under Article 4(11), Article 7 and Recital 32, consent must be freely given, specific, informed and unambiguous. Withdrawal must be as easy as giving consent, and access to a service must not be conditional on consent to data processing that is unnecessary for the performance of the contract. These requirements allow assessment not only of whether consent was formally given, but also of the conditions under which it was obtained. If an interface makes refusal significantly more difficult, conceals necessary information or directs users towards one option, the validity of consent under the GDPR may be questioned. Such a violation primarily concerns the lawfulness of data processing and does not automatically invalidate the contract.
A broader approach is established by Directive 2005/29/EC on Unfair Commercial Practices (UCPD). The Directive does not expressly prohibit dark patterns, but its provisions allow many such practices to be assessed by their effect on consumer behaviour. Article 6 of the UCPD covers misleading practices based not only on false information. A practice may also be misleading where factually correct information is presented in a way that is likely to deceive the average consumer and cause them to make a transactional decision they would not otherwise have made.
This is particularly relevant to digital interfaces, where manipulation may occur without false information. A trader may provide accurate information while visually emphasising one option, making another less noticeable, or requiring additional steps to select it. Although the content remains unchanged, its presentation may influence the user’s decision. Article 6 UCPD therefore permits the interface to be assessed as part of the commercial practice rather than merely as a technical means of implementing a decision already made.
Article 7 UCPD addresses misleading omissions. Material information must be presented so that consumers can make informed decisions. Accordingly, what matters is not only whether information is available on a website, but whether it is accessible when the decision is made. If material terms are hidden in additional menus, presented unclearly, or accessible only after several steps, the interface itself may affect consumer behaviour.
The restriction of consumers’ freedom of choice through pressure, coercion and undue influence is addressed by Articles 8 and 9 UCPD, which regulate aggressive commercial practices. Article 8 covers such influence where it significantly limits the consumer’s freedom of choice and causes them to make a decision they would not otherwise have made. Article 9 requires consideration of its circumstances, including duration, persistence and the creation of disproportionate barriers to exercising consumer rights, including the right to terminate a contract. This is particularly relevant to dark patterns that make cancellation significantly more difficult than acceptance.
This approach was further developed in Prezes Urzędu Ochrony Konkurencji i Konsumentów v Orange Polska SA (Case C-628/17). The President of the Polish Office of Competition and Consumer Protection challenged Orange Polska’s practice of requiring consumers to make a final decision to conclude a contract in the presence of a courier delivering the standard contractual terms. The Polish Supreme Court referred the case to the Court of Justice of the European Union for guidance on Article 8 UCPD. The Fifth Chamber held that such a procedure could not be regarded as an aggressive practice in every case. Such a finding requires consideration of all the circumstances and a determination of whether undue influence substantially limited the consumer’s freedom of choice. The Court therefore focused not on the form of contract formation itself, but on whether the particular procedure exerted pressure affecting the consumer’s decision. This reasoning is also relevant to dark patterns, as interfaces may exert pressure without direct coercion
More directly relevant to digital consent is Bundesverband der Verbraucherzentralen und Verbraucherverbände v Planet49 GmbH (Case C-673/17). The German Federal Court referred to the Court of Justice of the European Union the question of consent to the use of cookies expressed through a pre-ticked box. The Grand Chamber held that consent must be given through active user behaviour. A pre-ticked box therefore cannot demonstrate valid consent, as the user has not taken an independent action expressing their will. The Court also stressed the need to inform users about the duration of cookies and whether third parties could access them. This demonstrates that the mere ability to change a default setting does not establish free and legally valid consent. What matters is the structure of the choice and the action through which the user expresses their will.
The most direct regulation of dark patterns was introduced by the Digital Services Act (DSA). Article 25 of Regulation (EU) 2022/2065 prohibits online platforms from designing, organising or operating interfaces in ways that deceive or manipulate recipients of services or materially impair their ability to make free and informed decisions. Recital 67 identifies examples including unequal visual presentation of options, repeated requests to confirm a choice already made, making service cancellation more difficult than subscription, and default settings that are difficult to change.
The DSA therefore adds an important dimension to the regulatory framework. While the UCPD assesses manipulative interfaces through their effect on consumer behaviour and the GDPR regulates free and informed consent in personal data processing, the DSA directly addresses interface design and its impact on user autonomy. Article 25(2) excludes practices already covered by the UCPD or GDPR. The European model is therefore not a single prohibition, but a set of interconnected mechanisms applied according to the nature of the manipulation.
The European framework therefore supports an important distinction. A GDPR violation may call into question the validity of consent to personal data processing, while the UCPD addresses manipulation through deception or undue influence, and the DSA directly prohibits certain forms of manipulative interface design. For the validity of a particular transaction, however, it must be established whether the manipulation affected the formation of the user’s will, was material to the decision, and was causally connected to the transaction. Dark patterns should therefore be assessed according to their effects on the formation and expression of will.
Brenncke, who provides the first comprehensive mapping of EU legislative provisions expressly addressing dark patterns (including the DSA, the Data Act, the Consumer Rights Directive and the Consumer Credit Directive II), argues that these provisions share a common normative foundation. They protect biased consumers by adopting personal autonomy, rather than economic rationality, as the benchmark for assessment. On this view, regulating dark patterns in EU law is, in substance, regulating for consumer autonomy.
The European model also demonstrates that legal assessment may extend beyond the content of information to the architecture of choice. These instruments pursue different objectives: the GDPR focuses on consent to data processing, the UCPD on protection against unfair commercial practices, and the DSA on certain forms of manipulative platform design. They should therefore not be treated as interchangeable regimes. For contract law, their broader significance lies in recognising that the way information and alternatives are presented may be relevant to the legal assessment of digital choice.
The American Approach to the Regulation of Dark Patterns
In the United States, the regulation of dark patterns has developed differently from that of the European Union. There is no single federal statute imposing a general prohibition on such practices. The principal federal basis is § 5 of the Federal Trade Commission Act, which prohibits unfair or deceptive acts or practices. Dark patterns are therefore assessed through the concepts of deception and unfairness. The Federal Trade Commission (FTC) identifies as potentially unlawful practices the concealment of material terms, steering users towards a particular choice, making cancellation difficult, and interfaces that may lead to unwanted disclosure of personal data.
An important area of regulation concerns negative option transactions, where a user’s failure to cancel results in continued service or charges. The Restore Online Shoppers’ Confidence Act (ROSCA), adopted in 2010, requires disclosure of material terms, express informed consent before charging, and a simple cancellation mechanism. US regulation therefore addresses not only whether consent exists, but also how it is obtained.
This approach was further developed in FTC practice. In In re Epic Games, Inc, the Federal Trade Commission, relying on § 5 of the Federal Trade Commission Act, found that Fortnite’s interface could lead users to unintended in-game purchases. On 14 March 2023, the FTC approved a final order requiring $245 million in consumer refunds and prohibiting the use of dark patterns to charge users without affirmative consent.
A similar issue arose in Federal Trade Commission v Amazon.com, Inc. The FTC brought proceedings in the US District Court for the Western District of Washington, alleging that Amazon used manipulative elements in Amazon Prime subscriptions while making cancellation difficult. The Commission relied, among other provisions, on the Federal Trade Commission Act and Restore Online Shoppers' Confidence Act. Unlike Epic Games, the case involved judicial proceedings rather than an FTC administrative action. On 25 September 2025, the court approved a $2.5 billion settlement against Amazon.
These cases illustrate a feature of the federal approach. The FTC does not treat dark patterns as an independent basis of liability, but identifies the specific way an interface influences users and links it to existing prohibitions on deceptive or unfair practices. Legal significance therefore lies not in the design element itself, but in its effect on the user’s decision and the consequences of that effect.
More direct regulation has emerged at the state level. The most significant examples are California, Colorado and Connecticut, where dark patterns are linked to limits on user autonomy and independent choice. In California, this approach is reflected in the California Consumer Privacy Act (CCPA), which defines a dark pattern as an interface ‘designed or manipulated with the substantial effect of subverting or impairing user autonomy, decision making, or choice’ and provides that consent obtained through such practices is not valid consent in the relevant circumstances. Similar approaches are found in the Colorado Privacy Act and Connecticut Data Privacy Act, which also require free and informed consent.
Practical application of this approach can be seen in the work of the Connecticut Attorney General. In the Updated Enforcement Report Pursuant to Connecticut Data Privacy Act for 2024, the office identified the use of cookie banners and dark patterns that made consumer choice more difficult. In one case, the “AGREE” button was visually prominent, while refusal required navigating to an additional window. The Attorney General linked this practice to the principle of symmetrical choice, under which consent and refusal must be available simultaneously and must not differ substantially in difficulty or visual prominence.
The relevance of the American approach to contract law requires distinction. Federal regulation through deception and unfairness, together with state privacy laws, allows consideration of how users’ decisions are formed and, in some cases, treats consent obtained through dark patterns as invalid. These rules do not establish a general rule invalidating contracts concluded through dark patterns. Their significance here is that US law recognises the conditions in which consent is formed, not merely its external expression. This is relevant to whether a similar approach can be applied in Azerbaijan through the doctrines of vitiated consent and invalidity of transactions
Regulation of Manipulative Practices in the Republic of Azerbaijan
The Republic of Azerbaijan currently has no specific legal regulation of dark patterns. The term does not appear in existing legislation, and manipulative digital interface design is not recognised as a distinct subject of regulation. Nevertheless, existing provisions may apply to various forms of such practices. The key question is therefore whether the existing legal framework can address situations where interface design affects information perception, user choice, and the declaration of will.
The starting point for such an assessment is civil law. The Civil Code of the Republic of Azerbaijan, adopted in 1999, guarantees freedom in exercising civil rights and freedom of contract. Under Article 6.2, persons are free to exercise their civil rights and determine the terms of a contract. In the digital environment, a user’s decision is formed and expressed through an interface, raising the question of whether merely clicking a button is sufficient to constitute a legally significant choice.
Under Article 339.2 of the Civil Code, a transaction concluded as a result of fraud may be challenged where it would not have been concluded without such fraud. The provision therefore links fraud not only to wrongful conduct but also to its effect on the party’s decision, making it potentially applicable to certain dark patterns. Their use alone, however, is insufficient. It must be established what belief was created, whether it was material to the transaction, and whether the interface caused the transaction.
Contract law does not exhaust the legal approaches. Many dark patterns affect users before a contract is concluded, when they receive information about a product or service and choose between options. This makes the Law of the Republic of Azerbaijan ‘On Protection of Consumer Rights’, adopted in 1995, relevant. It establishes consumers’ right to receive reliable information about goods, works or services. In the digital environment, the availability and presentation of information matter. Where material information is hidden among other elements, accessible only after several steps, or visually less prominent than the option favoured by the seller, its availability may not ensure an informed choice.
This issue is directly relevant to electronic commerce. The Law of the Republic of Azerbaijan ‘On Electronic Commerce’, adopted in 2005, establishes specific requirements for electronic contracts. Under Article 7.3, electronic contract terms must be formulated so that a buyer without special knowledge can understand them clearly and unambiguously, while Article 8 requires certain information to be provided before an order is placed. Hidden additional charges, automatic inclusion of services, or interfaces that reveal material terms only at a later stage may therefore concern not only interface design but also statutory requirements for concluding electronic transactions.
Not every visual emphasis or alteration in the sequence of actions constitutes a legal violation. Where the user can readily understand the offer and independently choose an alternative, there is less basis for legal intervention. What matters is the degree of the interface’s influence and its effect on the particular decision
A similar limitation applies under advertising law. The Law of the Republic of Azerbaijan ‘On Advertising’, adopted in 2015, regulates advertising content and restricts information capable of misleading consumers. Accordingly, fictitious discounts, inaccurate prices, or false impressions of limited availability may fall within its scope. Its relevance is narrower where manipulation concerns only interface design or the sequence of actions, rather than the content of the advertising message.
Consent to the processing of personal data is particularly relevant. The Law of the Republic of Azerbaijan ‘On Personal Data’, adopted in 2010, sets requirements for obtaining data subject consent under Article 8.37 Under Article 8.2, consent must specify the recipient, processing purposes, data and processing operations, duration, and withdrawal conditions. Article 8.4 requires the owner or operator to prove that consent was obtained, while Article 8.6 provides for centralised consent and withdrawal through the ‘Electronic Government’ information system in cases prescribed by law.
This highlights a specific feature of Azerbaijani regulation. The law sets requirements for the content and evidential proof of consent, but neither defines dark patterns as a separate legal category nor renders consent obtained through manipulative interfaces automatically invalid. A violation of Article 8 therefore turns not on the unequal placement of 'agree' and 'decline' buttons as such, but on how the interface affected consent and whether the resulting consent complied with statutory requirements.
Overall, Azerbaijani law allows particular forms of dark patterns to be assessed through existing legal frameworks. Article 339 of the Civil Code applies where there is a defect in consent; consumer protection and e-commerce legislation addresses information and contractual certainty; the Law ‘On Advertising’ regulates misleading advertising practices; and the Law ‘On Personal Data’ establishes requirements for obtaining data subject consent.
Azerbaijani law primarily addresses the consequences of manipulative practices rather than the way in which choice is formed. Where a user is misled about a material contractual term or does not receive required information, existing legal mechanisms provide a basis for assessment. A gap arises where information is formally available and withdrawal remains possible, but the interface makes one option significantly more likely. Current legislation provides no independent criterion for such design. Azerbaijani law therefore regulates particular consequences of manipulation without treating interface design as an independent factor in legal assessment.
Conclusion
In light of this gap, the Law of the Republic of Azerbaijan ‘On Consumer Protection’ should introduce a specific provision addressing manipulative digital design. It should recognise dark patterns as an independent regulatory concern without replacing existing grounds for invalidity or making interface design automatically invalidate contracts.
The proposed provision could be formulated as follows:
A trader shall not use a digital interface that, through concealing material information, materially unequal presentation of options, preselection of additional services, or unjustified obstacles to refusal or cancellation, significantly and unreasonably restricts consumers’ ability to make free and informed decisions. Withdrawal, cancellation, and refusal must not be substantially more difficult than acceptance.
A central element of the reform should be the statutory classification of major dark patterns. As a basis, the OECD’s 2022 Dark Commercial Patterns classification should be used. It includes forced action, interface interference, nagging, obstruction, sneaking, social proof, and urgency, allowing dark patterns to be systematically classified by their impact on consumer choice.
In national legislation, forced action should cover unjustified demands for additional actions or excessive personal data. Interface interference should include visual emphasis of preferred options, preselected settings, and unequal presentation of alternatives. Nagging should cover repeated requests aimed at obtaining a decision favourable to the trader. Obstruction should encompass artificial difficulties in refusal or cancellation. Sneaking should cover hidden charges, services, or terms. Social proof should include misleading use of information about others’ behaviour. Urgency should cover false or unjustified impressions of scarcity or urgency.
This classification should remain open-ended. Timers, visual emphasis, repeated notifications, and preselected settings are lawful on their own and become legally significant only where they substantially and unjustifiably restrict consumers' ability to choose freely and informatively. This approach would allow the provision to address emerging forms of digital manipulation without constant legislative amendment.
A five-part functional test should determine whether a violation has occurred. First, identify the interface element or architecture capable of directing user behaviour. Second, assess the clarity, timing, and comparability of material information and alternatives. Third, determine whether the interface restricts free and informed choice. Fourth, assess the significance of its impact in light of the transaction and circumstances. Fifth, where the validity of the transaction or consent is disputed, establish a causal link between the manipulation and the resulting act.
The proposed framework should be incorporated into a new article of the Law ‘On Consumer Protection’. It should define dark patterns, provide an open list of their main categories, and prohibit substantial and unjustified restrictions on free and informed choice. For e-commerce, it should establish symmetrical choice, ensuring that refusal, cancellation, or changing a decision is not substantially harder than contracting or giving consent. In data protection, this principle would assess both consent itself and the conditions under which it is obtained, reflecting EU and US approaches to digital choice.
The legal consequences should operate at three levels. First, using a prohibited interface constitutes a regulatory violation and attracts statutory enforcement measures. Second, breaches of information or free-choice requirements provide consumers with appropriate remedies. Third, where a dark pattern causes a material mistake or fraud and causally induces the transaction, Article 339 of the Civil Code applies. Thus, an interface violation should not automatically render a contract invalid.
The proposed reform should not restrict ordinary UX design, personalisation, or product recommendations. It should target not persuasion itself, but substantial and unjustified interference with autonomous decision-making. Traders should remain free to use commercial design where material information is properly presented, alternatives remain accessible, and refusal does not become artificially more difficult.
The proposed model would combine the OECD classification, functional test, and existing civil-law remedies into a unified framework. Digital consent would be assessed not merely by the act of clicking, but by the conditions in which the decision was formed. This would shift Azerbaijani law from reactive regulation of manipulative outcomes towards preventive regulation of digital choice architecture, while preserving entrepreneurial freedom and avoiding automatic invalidity of digital contracts.
Bibliography
Primary Sources
Legislation and International/European Instruments
California Civil Code § 1798.140(h), (l).
Colorado Privacy Act, Colo Rev Stat § 6-1-1303(9).
Colorado Privacy Act Rules, 4 CCR 904-3.
Connecticut Data Privacy Act, Conn Gen Stat § 42-515 et seq.
Directive 2005/29/EC of the European Parliament and of the Council of 11 May 2005 concerning unfair business-to-consumer commercial practices in the internal market [2005] OJ L149/22.
Federal Trade Commission Act, 15 USC § 45(a)(1).
Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016 on the protection of natural persons with regard to the processing of personal data and on the free movement of such data (General Data Protection Regulation) [2016] OJ L119/1.
Regulation (EU) 2022/2065 of the European Parliament and of the Council of 19 October 2022 on a Single Market for Digital Services and amending Directive 2000/31/EC (Digital Services Act) [2022] OJ L277/1.
Restore Online Shoppers’ Confidence Act, 15 USC §§ 8401–8405.
Azərbaycan Respublikasının Mülki Məcəlləsi, 28 dekabr 1999, № 779-IQ.
Azərbaycan Respublikasının ‘Elektron ticarət haqqında’ Qanunu, 10 may 2005, № 908-IIQ.
Azərbaycan Respublikasının ‘Fərdi məlumatlar haqqında’ Qanunu, 11 may 2010, № 998- IIIQ.
Azərbaycan Respublikasının ‘İstehlakçıların hüquqlarının müdafiəsi haqqında’ Qanunu, 19 sentyabr 1995, № 1113.
Azərbaycan Respublikasının ‘Reklam haqqında’ Qanunu, 15 may 2015, № 1281-IVQ.
Cases
Bundesverband der Verbraucherzentralen und Verbraucherverbände – Verbraucherzentrale Bundesverband eV v Planet49 GmbH (Case C-673/17) EU:C:2019:801.
Federal Trade Commission v Amazon.com, Inc, No 2:23-cv-0932 (WD Wash, filed 21 June 2023).
In re Epic Games, Inc, FTC Matter No 2223087 (FTC, 14 March 2023).
Prezes Urzędu Ochrony Konkurencji i Konsumentów v Orange Polska SA (Case C-628/17) EU:C:2019:480.
Secondary Sources
Books and Articles
Cicero, Marcus Tullius, De Officiis III.58–60 (Walter Miller tr, Harvard University Press 1913).
Tom Blake, Sarah Moshary, Kane Sweeney and Steve Tadelis, ‘Price Salience and Product Choice’ (2021) 40(4) Marketing Science 619.
Martin Brenncke, 'Regulating Dark Patterns' (2024) 14(1) Notre Dame Journal of International & Comparative Law 39, 44, 78.
В Дождев, Римское частное право (Норма, Инфра-М 1996) 141–145
Егоров СА, ‘Противодействие негативному влиянию тёмных паттернов когнитивного дизайна в цифровой среде’ (2026) 3 Социально-гуманитарные знания 24.
Gray, Colin M and others, ‘The Dark (Patterns) Side of UX Design’ in Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems (ACM 2018) 1.
Mathur, Arunesh and others, ‘Dark Patterns at Scale: Findings from a Crawl of 11K Shopping Websites’ (2019) 3 Proceedings of the ACM on Human-Computer Interaction 81.
Суханов ЕА (ред), Гражданское право. Том I. Общая часть (3-е изд, Статут 2023) 457.
Thaler, RH, ‘Nudge, Not Sludge’ (2018) 9(2) Science 431.
Reports and Other Sources
Connecticut Office of the Attorney General, Updated Enforcement Report Pursuant to Connecticut Data Privacy Act (17 April 2025).
Federal Trade Commission, Bringing Dark Patterns to Light: Staff Report (September 2022).
OECD, Dark Commercial Patterns (OECD Digital Economy Papers No 336, OECD Publishing 2022).
Müəllif: Fidan Yusifli, Baku State University, Faculty of Law, 4th-year undergraduate student