When the Abyss Looks Back: Unveiling Evolving Dark Patterns in Cookie Consent Banners DOI: XXXXXXX.XXXXXXX Conference: Make sure to enter the correct conference title from your rights confirmation email; June 03–05, 2018; Woodstock, NY ISBN: 978-1-4503-XXXX-X/2018/06 CCS: Security and privacy Domain-specific security and privacy architectures

Nivedita Singh, Seyoung Jin, Hyoungshick Kim · 2026

Read When the Abyss Looks Back: Unveiling Evolving Dark Patterns in Cookie Consent Banners DOI: XXXXXXX.XXXXXXX Conference: Make sure to enter the correct conference title from your rights confirmation email; June 03–05, 2018; Woodstock, NY ISBN: 978-1-4503-XXXX-X/2018/06 CCS: Security and privacy Domain-specific security and privacy architectures by Nivedita Singh, Seyoung Jin, Hyoungshick Kim (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-21515.txt.

Introduction

Singh and co-authors study dark patterns in cookie consent banners, the tricks that steer users toward giving up data. As regulators tighten scrutiny, the tricks have grown subtler rather than vanishing. Their UMBRA system detects ten known patterns plus nine evolved ones, including pay-to-opt-out schemes and fake opt-outs. It mixes text analysis, visual checks, interaction tracing, and cookie monitoring. Scanning 14,000 websites, they found revocation often blocked, cookies set before consent, and opt-outs that fail to stop tracking.

Catalog record

CREATORNivedita Singh, Seyoung Jin, Hyoungshick Kim
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE79005
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.21515

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

When the Abyss Looks Back: Unveiling Evolving Dark Patterns in Cookie Consent Banners DOI: XXXXXXX.XXXXXXX Conference: Make sure to enter the correct conference title from your rights confirmation email; June 03–05, 2018; Woodstock, NY ISBN: 978-1-4503-XXXX-X/2018/06 CCS: Security and privacy Domain-specific security and privacy architectures Nivedita Singh, Seyoung Jin, Hyoungshick Kim Abstract . To comply with data protection regulations such as the EU’s General Data Protection Regulation (GDPR) and California Consumer Privacy Act (CCPA), websites widely deploy cookie consent banners to collect users’ privacy preferences. In practice, however, these interfaces frequently embed dark patterns that undermine informed and freely given consent. As regulatory scrutiny increases, dark patterns have not disappeared; but instead evolved into more subtle and legally ambiguous forms, making existing detection approaches increasingly outdated. We present UMBRA, a consent management platform (CMP)-agnostic system that detects both previously studied patterns (DP1–DP10) and nine newly evolved patterns (DP11-DP19) targeting information disclosure, consent revocation, and legal ambiguity, including pay-to-opt-out schemes, revocation barriers, and fake opt-outs. UMBRA combines text analysis, visual heuristics, interaction tracing, and cookie-state monitoring to capture multi-step consent flows that prior tools miss. We evaluate UMBRA on a manually annotated ground-truth dataset and achieve 99% detection accuracy. We further conduct a large-scale compliance-oriented measurement across 14,000 websites spanning the EU, the US, and top-ranked global domains. The results show that evolved dark patterns are pervasive: revocation is frequently obstructed, cookies are often set before consent or despite explicit rejection, and opt-out interfaces often fail to prevent third-party tracking. Beyond usability, we show concrete security implications: on sites exhibiting revocation barriers, cookies increase by 25% on average, and many cookies use insecure attributes that increase exposure to attacks such as XSS and CSRF. Overall, our findings provide evidence of systematic non-compliance and demonstrate how evolving consent manipulation erodes user autonomy while amplifying privacy and security risks. 1. Introduction Whenever users visit a website, a cookie consent banner or dialog typically appears, prompting them to set their preferences regarding third-party cookies. Obtaining consent for accepting or rejecting third-party cookies is valuable for websites, particularly for third-party advertising partners, and is required by GDPR ( 24 ) . The GDPR explicitly mandates that consent be freely given, informed, and unambiguous before personal data is processed, thereby placing cookie consent interfaces at the center of practical privacy enforcement ( 40 ) . As a result, cookie consent banners have become a primary mechanism through which regulatory intent is translated into real-world system behavior. However, many websites exploit this regulatory requirement by employing manipulative or coercive design practices, commonly referred to as dark patterns, that subtly nudge users into agreeing to data collection ( 21 ) . These tactics often make it difficult to opt out, for example by visually prioritizing the “accept” option, obscuring rejection controls, or presenting consent information in overly complex language. Such banners embody adversarial user interface strategies that systematically subvert user intent. Despite more than seven years of GDPR enforcement, these practices remain widespread. Regulators have imposed substantial fines for non-compliant consent mechanisms, including penalties of $170 million against Google and $68 million against Facebook ( 48 ) , yet cookie consent designs continue to evolve in ways that evade effective compliance. Recent regulatory guidance further requires websites to provide accessible and symmetric mechanisms for withdrawing consent after it has been given. In practice, however, similar dark patterns frequently persist in consent withdrawal interfaces ( 36 ) . These designs undermine user autonomy ( 29 ) and potentially violate the GDPR requirement, Article. 7 (3) that withdrawing consent must be as easy as giving it ( 13 ; 47 ) . Importantly, these revocation-related patterns do not merely affect user experience; they determine whether tracking continues despite an explicit attempt to opt out, directly shaping post-consent data collection behavior. Figure 1 . Examples of cookie consent banners featuring dark patterns, such as CookieInfoDisplay (DP11) and PurposeInfoDisplay (DP12) (Source: www.globenewswire.com ). Examples of cookie consent banners featuring dark patterns, such as CookieInfoDisplay (DP11) and PurposeInfoDisplay (DP12) (Source: \url{www.globenewswire.com}). Figure 2 . Example of an ideal cookie banner from our research perspective (Source: www.onetrust.com ). Example of an ideal cookie banner from our research perspective (Source: \url{www.onetrust.com}). Previous studies have taken important steps toward measuring dark patterns in cookie consent banners, primarily focusing on a fixed set of well-known patterns (DP1–DP10) ( 38 ; 7 ) . However, these approaches are increasingly outdated due to the rapid evolution of consent interfaces. Modern banners adopt more subtle and legally ambiguous strategies, such as multi-step consent revocation flows and pricing-based coercion mechanisms like “OptOutPricing (DP13)” and “Revocation Hard (DP15),” which are not captured by existing detectors. As a result, prior tools fail to identify a growing class of manipulative designs that emerge only through interaction and layered interfaces. Moreover, existing work largely treats dark patterns as usability or regulatory compliance issues and does not systematically examine their security and privacy consequences. In particular, prior studies do not analyze how manipulative consent designs affect cookie-setting behavior in practice. This omission is critical, as dark patterns can enable the deployment of excessive third-party cookies and trackers, thereby increasing exposure to security risks arising from insecure cookie policies ( 36 ) . Without linking interface-level manipulation to concrete system behavior, the real-world impact of evolved dark patterns remains poorly understood. To address these gaps, we present UMBRA, a Consent Management Platform (CMP)-agnostic system detecting both DP1–DP10 and nine newly evolved patterns (DP11–DP19), including revocation barriers and fake opt-outs. UMBRA combines text analysis, visual heuristics, interaction-based monitoring, and cookie-state analysis to capture complex consent flows that manifest only after user interaction. Using UMBRA, we conduct the first compliance-oriented, large-scale measurement study of evolved dark patterns, guided by the following research questions: RQ1. Can a CMP-agnostic detector accurately identify both established (DP1–DP10) and newly evolved dark patterns (DP11–DP19) across diverse cookie consent banners? RQ2. How widespread are newly evolved dark patterns (DP11–DP19) across different regions and website categories? We summarize our contributions as follows: • Enhanced detection of evolved dark patterns. We extend existing detectors by incorporating nine newly evolved dark patterns (DP11–DP19) that capture contemporary consent manipulation strategies, including revocation barriers and pricing-based coercion. • Security-aware cookie-state analysis. Beyond interface detection, UMBRA monitors cookies across different interaction states. We show that websites exhibiting revocation Hard (DP15) set 25% more cookies on average, many of which increase security and privacy risks. • CMP-agnostic detection framework. We design a CMP-agnostic framework that integrates dialog extraction, clickable classification, text and visual analysis, and cookie-state monitoring. • High detection accuracy at scale. We evaluate UMBRA on a manually annotated dataset, achieving 99% accuracy, and conduct a large-scale measurement across 14,000 websites from the EU, the US, and top-ranked global domains. Prior studies reported accuracy of 91% ( 24 ) . • Evidence of widespread non-compliance. Our study provides the first comprehensive compliance-oriented evidence that evolved dark patterns are pervasive and closely linked to increased tracking and insecure cookie practices. Overall, our findings demonstrate that consent manipulation remains systematic and increasingly sophisticated, eroding user autonomy while amplifying tangible security and privacy risks. By bridging interface-level detection with security-aware measurement, this work provides actionable insights for researchers, regulators, and practitioners seeking to enforce

Read the complete text