Discord Begins Age Assurance Checks for All Users
Discord will begin applying age-assurance measures to all users starting Wednesday, September 23, 2026. TechCrunch The company has already launched age assurance in Brazil and Texas, according to CTO Stanislav Vishnevskiy. Confirmed adult accounts will see no change to the platform. Most users will notice nothing.
Discord delayed its plans in February 2026 after user backlash over concerns that biometric scans and ID checks could jeopardize privacy. That concern was sharpened by a 2025 incident in which a third-party vendor working with Discord was breached, potentially exposing data of at least 70,000 users including government ID photos and selfies used for identity verification. The earlier plan would have required verification for full access, with only users verified as adults able to change certain settings. TechCrunch
Under the revised system, Discord says more than 90% of users will not have to verify their age. The company will use account tenure, device and activity data, and other usage patterns to classify users into adult, teen, or unconfirmed age groups. It uses a machine learning model trained on reliably confirmed data linking user behavior to age to predict whether an account likely belongs to an adult or a teen. Discord states it does not estimate users' ages based on message content, profile data, or demographic attributes.
The distinction matters for implementation. Behavioral inference becomes the default classifier, with explicit verification reserved for edge cases where confidence is low. Users not confidently identified as adults can prove their age without a biometric or ID scan by sharing a credit card or age group data from their Apple App Store or Google Play Store account. The fallback matters. It shifts trust to payment rails and mobile OS age signals rather than storing face maps and ID images on Discord-adjacent systems.
Discord said it delayed the global rollout until the second half of 2026 to expand verification options. Discord The company launched teen-by-default settings globally and said over 90% of users globally will be assigned new default settings that support age-appropriate experiences while keeping privacy front and center. Teen accounts will be automatically blocked from accessing age-restricted servers or channels. Message requests from non-friends will be routed to a separate inbox, and teens will receive alerts when accepting friend requests from someone with no mutual friends.
Discord has documented the changes in a Policy and Safety blog post titled "Getting Global Age Assurance Right: What We Got Wrong and What's Changing" and in an Age Assurance Update FAQ on its official support site. The February delay and the revised verification menu suggest product and policy teams were responding directly to the criticism of document-first verification, while retaining stricter defaults for accounts classified as teen or left unconfirmed.
In this author's view, the architecture reflects a pragmatic compromise familiar to trust and safety engineering. Document checks provide high assurance per user but create centralized stores of sensitive credentials, expand third-party risk, and introduce friction that drives evasion. Behavioral classification inverts that tradeoff. It lowers friction for the large adult majority at the cost of probabilistic errors at the margins, including false positives that push legitimate adults into manual review and false negatives that leave some teens under-protected. Worth flagging for practitioners is the explicit scoping of inputs, tenure and device and activity signals only, which constrains function creep but also limits signal quality compared with content-aware models.
Looking at what this enables if the classifier performs as described, Discord preserves open onboarding while applying age-appropriate defaults by default rather than by opt-in. Teen-by-default routing of stranger messages and blocks on age-restricted spaces reduce exposure without requiring every user to upload identity evidence. The long-term gain is a verification stack where IDs and biometrics are the exception, not the entry ticket, and where platform safety work can iterate on signals and thresholds instead of re-litigating document collection with each regulatory shift.
