CATALOGDARK PATTERN
LABSOURCE taxonomy (opens in a new tab)
Information Manipulation

Feedforward Ambiguity

Feedforward Ambiguity is an interaction-based dark pattern formally categorized under “(De)Contextualizing Cues” and “Interface Interference.” While feedback informs a user about what has happened, feedforward informs the user about what will happen before they initiate an action. This pattern deliberately provides unclear, polysemous, or entirely missing cues regarding the consequence of an interaction. Common implementations include vague button labels (e.g., “Continue” to finalize a non-refundable purchase), icon-only buttons lacking accessible labels, or toggles that do not clearly indicate their activation state. This forces the user to guess the outcome, effectively inducing erroneous compliance with unintended financial or privacy commitments through a lack of functional transparency.

1

Structural Ambiguity of Action-Outcome Mapping

Feedforward Ambiguity: Structural Ambiguity of Action-Outcome Mapping

Order review

Step 2 of 3 — one more step after this one.

Order summary
Studio headphones$59.00
Carry case$25.00
ShippingFree
Total$89.00

Non-refundable once placed.

Condition 1: Structural Ambiguity of Action-Outcome Mapping
Given
To establish a structural baseline for Feedforward Ambiguity, the algorithm inspects interactive elements whose onclick, href, or formaction targets resolve to URLs or state transitions with semantics that conflict with the element's rendered label. Let be the user's expected outcome derived from the button label, and be the actual structural target. The feature triggers if the semantic divergence between expectation and resolution exceeds a threshold :
2

Iconographic Entropy and Missing Affordances

Feedforward Ambiguity: Iconographic Entropy and Missing Affordances

Security settings

Manage your account security and sharing preferences.

Account protection
Activity alerts
Partner integrations

Choose an option to continue.

Condition 2: Iconographic Entropy and Missing Affordances
Given
Visual polysemy is frequently weaponized by removing textual anchors from interactive icons. We define as a graphical interactive node that lacks adjacent visible text and accessibility attributes. Using an icon-classification model, we derive , representing the probability distribution of possible semantic meanings. We calculate the Shannon entropy of these predictions, where high entropy indicates the icon is contextually ambiguous. The feature triggers if the interface relies on a highly ambiguous icon for a critical action without providing a clarifying tooltip or accessible label:
3

Semantic Divergence of Action and Outcome

Feedforward Ambiguity: Semantic Divergence of Action and Outcome

Create your account

Step 1 of 3 — account details.

Condition 3: Semantic Divergence of Action and Outcome
Given
To identify deceptive labeling, we define as the text label of an interactive DOM node (e.g., “Next” or “I Agree”). We map the user's expected outcome to a semantic vector space, , and contrast it with , the actual backend execution or state transition triggered by the interaction (e.g., SubmitPayment() or OptInAll()). The feature triggers if the semantic similarity between the predicted linguistic intent and the actual programmatic outcome falls below a clarity threshold . This is particularly critical if the outcome resides in a high-stakes domain , such as finance or data privacy: