Consent capacity stratifies with age, not acquired in a single day

What it is

The capacity to understand "what am I consenting to, what follows, and that I may refuse" develops continuously, not on a single birthday. Graduated consent capacity therefore argues for age-stratified consent mechanisms: for young children, consent is exercised by a guardian while the child gives developmentally matched assent (informed in comprehensible terms and given a chance to agree); with age, decision rights transfer incrementally — low-stakes decisions first (avatar, friend requests), then medium-stakes (new features, content visibility), with high-stakes and data decisions transferring last. Medicine's "mature minor" handling is the precedent for this structure.

Why it happens

A binary adult/child cutoff distorts in both directions. Downward: granting consent to a child who cannot understand empties the act — a "tap on agree" with no comprehension behind it is a button press, not consent; it also strips the guardian's proxy role of its legal and ethical grounding. Upward: adolescents near adulthood already understand most everyday decisions well, and full proxy control creates needless friction while removing the practice of deciding — autonomy develops precisely through real decisions, and blanket guardianship merely postpones the capability. Stratification's mechanical value is aligning the comprehension threshold of consent with the risk level of the decision: each risk tier opens to personal consent only at the age where capacity matches, with guardianship and assent covering the rest.

Where it stops holding

Stratification needs an operational age signal, which loops back to the privacy trade-off of age verification — the finer the tiers, the heavier the verification burden, so practice compromises at two or three tiers rather than a continuum. Statutory consent ages and tier structures differ by jurisdiction (data-consent thresholds range from thirteen to sixteen); local law is the floor, and ethical tiering may be finer but never coarser. Development is also individually variable: same-age comprehension spreads widely, so age tiers are a cost–precision compromise, not developmental fact itself.

Applying it

• Define a consent-tier table: decision category × risk level × consent mode (guardian-only / guardian consent + child assent / self-consent with guardian notice / self-consent), and put it in the design spec.
• Build the assent step from age-comprehensible material: one thing at a time, consequences instead of clauses, refusal allowed and penalty-free; a failed assent is not forced through.
• Set automatic transfer points as age thresholds are crossed: proactively announce "these decisions are now yours" rather than waiting to be discovered; transfers have a transition phase (self-consent plus guardian notice).
• Verify: sample-interview each age tier about what was just consented (what, consequences, revocable?); a tier that cannot answer has a failed consent design — demote its mode.

Related

• Same group: P4.05.3 Age verification itself involves a privacy trade-off · P4.05.7 Parent-facing explanations cannot replace child-facing expression
• Adjacent: O2.06 Consent interface design · P4.05.1 Defaults must be set for the youngest plausible user
• Search terms: assent vs consent · graduated autonomy · mature minor