Explicit Risk Communication
Proactively flag concerning clauses so users can assess risks without having to parse legal language themselves.
Design Guidelines
Identify and highlight high-risk clauses such as third-party data sharing, indefinite retention periods, and broad usage rights using clear visual indicators.
Consider using a color-coded system to signal concern levels. For example: red for high concern practices that significantly impact privacy, yellow for clauses requiring careful consideration, and green for standard low-risk provisions.
Provide plain-language explanations of why flagged clauses matter and what they mean for users in practical terms.
Surface these risks early in the consent flow rather than burying them within full legal text, and allow users to filter or view only high-risk items for efficient assessment.
Do's and Don'ts
Don’t
Hide concerning practices within dense legal paragraphs
Treat all clauses as equally important with no differentiation
Require users to read everything to discover concerning practices
Use vague language like "trusted partners" for data recipients
Do
Use color-coded visual indicators to flag risk levels
Highlight high-risk clauses prominently with clear explanations
Provide "View high-risk clauses only" filter for quick assessment
Specify exactly what data is shared and why it poses risks to the user
Research Foundation
Research revealed that high comprehension did not necessarily signal high trust. Total Consent achieved the highest comprehension scores (Mean=3.55/5) but received the lowest trust ratings (Mean=3.1/7), with with 40% of participants declining terms after understanding the full extent of data collection. One participant stated emphatically upon seeing password requirements:
"Passwords is a no no! I don't care if I can't use the app" (P04).
The participant immediately declined , despite rating clarity high. This suggests users need both comprehension AND risk assessment simultaneously as opposed to the wide practice where designs force them to choose one or the other (Richards & Hartzog, 2019).