Beyond Dark Patterns: A Concept-Based Framework for Ethical Software Design

The paper addresses the gap between dark pattern taxonomies (which emphasize what designers should avoid) and practical guidance on what designers should do ethically. The research question asks whether a concept-based framework—grounded in positive, standard units of functionality (software concepts)—can define expected behavior, evaluate deviations as dark, and provide actionable guidance for design and regulation. The context is widespread user resignation to deceptive design and limitations of negative catalogs in handling variants and completeness. The purpose is to define acceptable designs as instantiations of positive, standardized concepts, and to judge darkness as deviation from user expectations that benefits the provider at the user’s expense. The study’s importance lies in offering a language for consensus standards, a method for design via concept catalogs, and an evaluative framework that is both procedural (deviation from standard concepts) and substantive (harm to users), aiming to bridge academic insights and industry practice.

The paper reviews the evolution of dark patterns since Brignull’s introduction of the term, covering taxonomies and strategies (e.g., nagging, obstruction, sneaking, interface interference, forced action) and theoretical efforts to define darkness (Mathur et al., Gray et al.). Prior work often relies on modifications to default choice architecture and underlying psychological mechanisms to explain manipulation. The authors argue these approaches leave ambiguity regarding the default architecture and user expectations. They situate their framework by emphasizing expected behavior via standardized concepts, aligning with users’ felt experience (affordances and anticipated outcomes), but making deviation from expectations primary rather than secondary to strategies. The review also connects to ethical theories distinguishing procedural and substantive criteria, and to debates around ethical nudges, transparency, salience, and force. Related work on user expectations notes context-dependence and the co-evolution of designer practices and user familiarity. The authors propose concept catalogs to formalize conventional practice and expectations, complementing platform-specific guidelines and UI pattern libraries by focusing on functionality rather than widgets.

The methodology comprises developing a concept-based framework and evaluating it through three studies, alongside constructing a sample concept catalog entry. Framework development: The authors formalize software concepts as abstract, reusable units of functionality with defined state, actions, synchronizations, and UI mappings. They define three concept perspectives—observed (user’s interpretation via UI), implemented (actual behavior), and expected (user’s prior experience)—and specify darkness when observed or implemented concepts deviate from expected concepts and the deviation benefits the provider to the user’s detriment. They refine the definition using standard concepts from catalogs. Concept catalog construction: A standard Shopping Cart concept is defined with minimal core functionality (states: owner, items, quantity, price, subtotal; actions: createCart, add, changeQuantity, remove, changePrice, checkout; synchronizations with Catalog, Order; mapping guidance for subtotal and changeQuantity), plus general mapping principles (transparency, correspondence, faithfulness, consistency, symmetry, conventions). Study RQ1 (coverage of existing dark patterns): They re-described each entry in Brignull’s updated Deceptive Patterns taxonomy (April 2023) using concepts, mapping each pattern to relevant concepts and identifying violations (e.g., fake scarcity as mapping transparency failure in Inventory state; forced action as unexpected synchronization between Account.create and Contact.share). Study RQ2 (nuanced evaluation): Case study of three sites with shopping behavior variants—Sports Direct (sneaking: auto-added magazine), Grove Collaborative (suggested items auto-added; recurring auto-ship; transparent UI and reminders), StitchFix (personal shopper-like flow; stylist-selected shipments; subscription by default). They analyze user reviews (e.g., Grove has mixed perceptions; StitchFix reviews ~200 with no accusations of deceptive design) to triangulate user sentiment and evaluate each design via concept deviation and harm. Study RQ3 (common functionality patterns): One researcher built the Shopping Cart catalog entry by surveying major platforms (WooCommerce, Squarespace, Shopify) and leading sites (Amazon, Etsy) and reviewing blogs and academic literature on e-commerce carts; another researcher manually examined the top 20 US e-commerce sites by Similarweb (June 2023), excluding non-English or sites lacking shopping carts, to test whether the standard concept covers core functionality. They documented common extensions (e.g., multiple quantity add, save for later, partial checkout) and assessed conformity to the core concept.

• The concept-based framework can describe all dark patterns in Brignull’s taxonomy by identifying deviations from expected or standard concepts, often via UI mapping misrepresentation or unexpected synchronizations. This answers RQ1 affirmatively and adds nuance (e.g., distinguishing explicit misrepresentation from subtler expectation violations). - The framework evaluates nuanced cases beyond taxonomy labels (RQ2). Sports Direct clearly violates the Shopping Cart concept via system-initiated add() (sneaking). Grove Collaborative presents a Shopping Cart but allows shop-initiated add(), transparently communicated; users’ expected concept often presumes user-only add(), yielding mixed perceptions of darkness. StitchFix communicates a distinct Personal Shopper concept, aligning observed and expected concepts and avoiding perceived sneaking despite similar automated behaviors. The framework suggests design fixes: restrict add() to user, or reframe functionality under a distinct concept (Personal Shopper) or as a Recommendation concept alongside the cart. - Common functionality across websites is consistent enough to be captured succinctly by a standard concept (RQ3). All 20 top US e-commerce sites examined conformed to the core Shopping Cart concept definition and mapping standards, while most implemented extensions. The catalog entry required no changes to describe core functionality across sites, supporting feasibility of concept catalogs and their reusability. - Additional observations: Among approximately 200 StitchFix reviews sampled, none accused deceptive design, aligning with proper concept communication. Concept analysis highlights darkness at both semantic and UI mapping levels and provides a structured language for standards and regulation.

Findings show that a concept-based approach offers a principled and practical lens for evaluating dark patterns. By grounding darkness in deviation from standard or expected concepts that harms users, the framework accounts for existing patterns, clarifies nuanced cases where user perceptions diverge, and enables generative guidance for redesign. The approach contributes: a method to define user expectations via documented standard concepts rather than implied choice architectures; coupling procedural criteria (deviation from standard functionality/mapping) with substantive evaluation of user harm; recognition of darkness both in underlying semantics and in UI representation; and provision of a structure (concept catalogs) for consensus standards usable by designers, companies, and regulators. The case studies illustrate how correct concept communication (e.g., Personal Shopper vs. Shopping Cart) aligns expectations and mitigates perceived darkness. The catalog feasibility study demonstrates broad reuse of identical functional units, supporting standardization and industry alignment. The discussion also acknowledges the challenges of culture- and context-dependent expectations, non-uniform perceptions of harm, and the risk that standardization could entrench problematic concepts, suggesting ongoing debate and refinement within catalogs.

The paper proposes a concept-based framework that complements existing dark pattern research by focusing on positive, standardized units of functionality and mapping deviations that harm users. It shows the framework can explain established dark patterns, evaluate nuanced designs, and that concept catalogs can succinctly capture common functionality (e.g., Shopping Cart) across many sites. The main contributions are: a framework for evaluating designs via concept deviation and user detriment; a language and structure for consensus standards through concept catalogs; and a design method that separates catalog development from application instantiation. Future work includes field studies with practicing designers to assess adoption and usability, development of broader concept catalogs (possibly by industry consortia or governmental agencies), exploration of culturally and contextually specific variants, and refinement of standards to align with societal normative values while maintaining practical applicability.

• No practitioner evaluation: The studies did not involve practicing designers using the framework or catalogs, limiting insights into practical adoption and usability. - Limited scope of RQ2: The case study focuses on shopping cart-related functionality (adding items, recurring shipments), potentially missing other dark patterns present in the applications. - Single-concept evaluation in RQ3: Only the Shopping Cart concept was evaluated across sites; other concepts may be less uniformly implemented, constraining catalog utility. - No comparison to underlying psychological techniques: The framework is not evaluated against specific persuasive or manipulative strategies, which may capture nuances the concept approach omits. - Culture- and domain-dependence: User expectations and standard concepts may vary by culture or domain, complicating universal standards. - Non-uniform perceptions of harm: Substantive evaluations depend on user values and context, making some judgments inherently subjective. - Standardization risk: Making a concept standard may legitimize problematic designs (e.g., Newsfeed), requiring explicit debate and normative guidance in catalogs. - Novel designs: Completely new concepts can evade indictment until expectations form, creating a window where manipulative elements are difficult to judge.