Discord Claims Over 90% Users Do Not Need Age Verification
Stanislav Vishnevskiy, co-founder and CTO of Discord, stated that over 90% of users do not need to verify their age. Starting this week, the platform will categorize accounts into adult, teen, and unverified tiers, using signals such as account age, device and activity, and community membership to make assessments, without reading message content, personal profiles, or demographic attributes.
The adult experience remains unchanged. Teen and unverified accounts can still retain their accounts, servers, friends, private messages, and voice features, but friend requests from strangers will go to a pending list, sensitive images will be blurred or intercepted, access to age-restricted channels will be limited, and some default safety settings cannot be changed. The company noted that the number of communities and interaction methods for typical 15-year-old and 30-year-old users differ at an aggregate level.
After opposition to biometric and ID verification in February, the global rollout was postponed from March to the second half of 2026. At that time, unverified users were defaulted to the teen experience. It has now been piloted in regions with age verification legislation, such as Brazil and Texas, while continuing to comply with mandatory verification under the UK Online Safety Act.
The approximately 10% of users who need verification can choose from device-based facial age estimation, ID, credit cards, age ranges from Apple and Google app stores, Google Wallet, and reusable credentials like AgeKey. The company requires facial age estimation to be completed on the device, ensuring biometric data does not leave the phone, having previously abandoned the supplier Persona for not meeting standards and switching to Incode. Users who do not verify will lose access to age-restricted content and the ability to modify certain safety settings, while retaining other functionalities.
Those buying are looking to write enforceable legislation and app store policies to protect minors, while those selling are using behavioral signals to replace universal ID verification on the platform. Funding has shifted from verification suppliers to in-house security models. Beneficiaries are users who do not access restricted content and have sufficiently long account histories, while those under pressure are new accounts wanting to access restricted channels but refusing to provide ID, as well as adult users defaulted to the teen tier.
Technical blog posts will simultaneously disclose the logic behind the age estimation model. The rollout will gradually cover all users within a week.
Source: Public Information
ABAB AI Insight
Discord grew through gaming voice and private servers, with users treating anonymity as the product itself. The February plan to default everyone to teen status and force ID or facial recognition essentially dismantled anonymity using regulatory language. After Vishnevskiy publicly acknowledged the mistake, the approach changed to "signal bucketing + minimal verification," translating the UK Online Safety Act, Brazil's Digital Child Law, and Texas regulations into a deployable engineering solution, rather than engaging in another privacy public relations battle. The abandonment of Persona and the switch to Incode for device-based age estimation indicate that supplier choices have been formalized into compliance procurement standards.
The capital path is to replace incremental identity collection with existing behavioral data. Account age, device, and server entry records are byproducts of anti-spam systems, now re-labeled as age characteristics. The 90% exemption from verification reduces customer service demands and churn, while the 10% verification targets high-risk actions related to accessing restricted content and modifying safety settings. Credit cards and app store age ranges shift verification costs to payment networks and Apple/Google, with Discord only collecting age results, not identities.
This is comparable to YouTube treating non-logged-in users as minors, Meta requiring ID verification by region, and Steam locking content by store tier. The industry phase has shifted from "platform self-governance community standards" to "legislative mandatory age assurance": first piloting in jurisdictions with laws, then rolling out the same bucketing system globally. A failed comparison is the communication breakdown in February—users believed everyone was required to provide facial recognition, while the company claimed it was never designed that way.
Structural changes are a response to regulatory changes. Once age becomes the ticket to access restricted content, platforms must choose a verifiable classifier between privacy lawsuits and fines. The mechanism is: the law requires provable age assurance, while users want to avoid ID submission; behavioral signals allow platforms to claim they "know" the age of most users, charging verification only for those needing to access restricted areas. Whoever controls the bucketing model decides who is treated as a minor.
ABAB News · Cognitive Law
Ninety percent exempt from verification shifts costs to that ten percent's doorstep.
Once behavioral signals are used as age indicators, anonymity begins to stratify.
Regulators want proof, while users want to avoid ID submission.