Autonomy
Interfaces should support deliberate choices rather than exploit confusion, friction or behavioural vulnerabilities. Remediation should make refusal, cancellation and withdrawal genuinely usable.
Consumer & child protection IWIB4AI session Needs review
WISE × IWIB4AI – Dark Patterns, Digital Harms & AI in Compliance
Using AI to detect manipulative interface design and support consumer protection compliance and remediation.
Draft – legal and case law references not yet verified
Summary
AI-assisted tools can help compliance and design teams identify potentially manipulative interface patterns and prioritise remediation. These patterns can steer people towards unwanted purchases, subscriptions or disclosures of personal data, creating consumer harm and legal and financial exposure. Effective use combines automated detection with contextual legal assessment and human-led redesign.
The Problem
Interfaces can serve business goals while undermining users’ ability to make informed, voluntary choices. AI may help identify these risks at scale, but interface features alone do not establish whether a design is unlawful: context, effects and the applicable legal framework matter.
Ethical Grounds
Fairness
Manipulative design can disproportionately affect children and people with limited digital literacy or other vulnerabilities. Audits should examine how different user groups experience the same interface.
Transparency
Prices, recurring commitments and data-use choices should be presented clearly before users act. AI-generated audit findings should identify the design feature and explain the suspected harm.
Accountability
Organisations remain responsible for interface design even when AI supports compliance checks. Findings need named reviewers, documented decisions and verification that remediation addresses the underlying harm.
Legal Compliance
| Framework | Provision | Relevance |
|---|---|---|
| DSA | Art. 25 | Article 25 prohibits online platforms from designing, organising or operating interfaces that deceive or manipulate users or materially impair free and informed decisions, subject to its exclusion for practices covered by the Unfair Commercial Practices Directive or GDPR. |
| EU consumer law | Directive 2005/29/EC, Arts. 5–9 | Consumer-facing dark patterns may constitute unfair, misleading or aggressive commercial practices, requiring assessment of their context and effects on consumers’ transactional decisions. |
| GDPR | Arts. 4(11), 5(1)(a) and 7 | Where interfaces concern personal-data processing, manipulative design may undermine fairness and transparency or invalidate consent, and withdrawing consent must be as easy as giving it. |
Sources
Court & Regulator Decisions
Bundesverband der Verbraucherzentralen v Planet49 (opens in a new tab)
The rejection of pre-ticked cookie consent supplies a concrete legal benchmark for detecting interfaces that substitute defaults for informed, active choice.
Key Takeaways
- Use AI findings to flag risks for human review, not as automatic legal conclusions.
- Audit complete user journeys, including refusal, cancellation and consent withdrawal.
- Map each suspected pattern to the applicable law and assess impacts on vulnerable users.
- Assign remediation owners and test whether redesigned interfaces enable informed, voluntary choices.
Related Use Cases
From the Session
WISE × IWIB4AI – Dark Patterns, Digital Harms & AI in Compliance
Speakers
