Opt-in vs. opt-out
Opt-in waits for a clear yes before processing; opt-out processes until a person says no. The law decides which model fits each use.
What it is
Opt-in means the processing stays off until a person takes a clear affirmative action. Opt-out means it starts or remains on until the person refuses. The right model depends on the activity and law, not on which toggle converts better: EU non-essential device storage generally needs opt-in consent, US CAN-SPAM commercial email uses an opt-out model, and California gives people an opt-out from the sale or sharing of personal information.
Decide per purpose, channel, person, and jurisdiction, then make the interface and backend express the same state. Opt-in needs an empty default and proof of the affirmative act. Opt-out needs a visible, effective refusal route and processing that actually stops when it is used.
Gotcha: silence, inactivity, and pre-ticked boxes are not valid consent. A switch labelled 'Disable tracking' also reverses the meaning of its stored boolean and breeds bugs. Store positive purpose states, the legal model, notice version, timestamp, and source instead of one mysterious marketingAllowed flag.
Ask AI for it
Build an opt-in versus opt-out decision matrix and consent implementation for this product. Separate every purpose and channel, then map EU ePrivacy consent, GDPR lawful basis, CAN-SPAM unsubscribe, California's Do Not Sell or Share right, and the Global Privacy Control Sec-GPC: 1 header. For each row specify the default state, affirmative action or refusal route, evidence stored, withdrawal deadline, and downstream systems to stop. Implement explicit purpose states and map them to Google Consent Mode v2 signals analytics_storage, ad_storage, ad_user_data, and ad_personalization. Include tests for first visit, refusal, withdrawal, policy-version change, and a recognized GPC signal.