A major platform requires proof of personhood or bans AI content

Where things stand (17 Oct 2026)

As far as I can find, no social or content platform with at least 100 million monthly users meets either branch, and none has announced a policy that would. The industry is clearly moving toward checking who is human, but every step so far falls into one of the question's exclusions:

• Reddit (25 Mar 2026) asks only "fishy" accounts to prove they are human, and says outright "this is not sitewide human verification, let alone sitewide ID verification." Passkeys are its preferred method, and passkeys alone don't count. (Biometric Update)
• X (Oct 2026) sends liveness checks only to flagged accounts. Accounts that skip it lose posting, following and direct messages, not access to the account. At the same time X is inviting brands to run Grok-powered bot accounts. (MediaPost, Social Media Today)
• Facebook Verified (24 Jul 2026) is a free, optional selfie badge. The head of Facebook says Meta hasn't decided whether Facebook should be human-only. (Meta, Sources)
• Roblox requires age checks to use chat, and unverified users are pushed into a kids-style experience. That is a feature restriction, and it still contributed to an 18% one-day stock drop when guidance was cut. (Engadget, shattered.io)
• Discord (Sep 2026) estimates more than 90% of users won't be asked to verify. (Ars Technica)
• Tinder comes closest in policy terms: a face check is a "condition of being on the service," now extending to existing users. But it applies only in some regions, and Tinder has about 60 million users. (WIRED, Digital Watch)
• Government ID mandates are all country-specific, which the question excludes:
  • Turkey's ID-login plan (Nordic Monitor).
  • Malaysia's ID-based age checks.
  • Meta's US settlement requiring age checks (Tech Policy Press).
  • China's national online ID is voluntary, with about 40 million sign-ups (chinanews). Its draft rules on minors (comment period closed 17 Oct 2026) don't require ID checks for every account (Beijing News).
• New human-only networks:
  • OpenAI's biometric social network was reported in January 2026 as an early project with fewer than 10 staff, and there has been no launch news since (Gizmodo/Forbes).
  • World has about 18 million iris-verified people out of 39 million users (Stocktwits).
  • The European network W lets unverified accounts read; only posting requires verification (Cybernews).
• AI-content bans:
  • English Wikipedia's March 2026 ban on AI-written text predates the window and allows AI help with copyediting and translation (How-To Geek).
  • Snapchat bans only fully AI-made videos from Spotlight, and Snapchat supplies its own AI tools (Hollywood Reporter).
  • Among music services, only Bandcamp bans AI music, and it is far below 100 million users (Billboard).

Base rate and paths

Base rate. The comparison class is drastic universal gates imposed by 100-million-plus platforms. No such platform has ever required every account worldwide to pass an ID or biometric check or lose access; China's phone-based real-name system is excluded by the question. Musk has said since 2022 that X would "authenticate all humans," and it still hasn't. Ideas like X going paid-only were floated and never rolled out globally. The historical base rate is about zero. I adjust upward because of the five-year horizon, the growing pressure from AI agents, and the verification tools now being built.

Paths to YES and my rough probabilities through 2031:

Path	Probability
A new verified-only network (e.g. OpenAI's) reaching 100 million monthly users	~1.3%
An existing Western platform (X, LinkedIn, Facebook, etc.) making verification universal, with loss of access	~1.2%
Chinese regulators requiring ID or face checks for every account, in a way that counts	~1.2%
Other national mandates on domestic-only platforms, or universal age checks being counted	~0.5%
The AI-content ban branch (blocked by exceptions for educational examples and platforms' own AI features)	~0.2%

Allowing for overlap, that comes to about 4%.

Shape over time. I modelled a quarterly chance starting at 0.08% for Q4 2026, since nothing is scheduled and a policy would need to be announced and have its transition finished within 2.5 months. It rises to about 0.2% per quarter by 2028 and levels off at 0.23%. Multiplying through gives a cumulative 0.6% by end-2027, 1.4% by end-2028, and 4.0% by end-2031.

Strongest case for higher. Verification is becoming a condition of participation in dating apps, and Meta, Google and X now have selfie-check tools in place. One platform decision could flip this. A Manifold market ("major platform implements privacy-preserving proof of personhood by 2030", about 55%) uses a much looser standard, so it doesn't move me much.

Strongest case for lower. The question's own exclusions (optional, selective, regional, new-accounts-only, feature-restriction) cover every real-world move so far. Platforms have strong reasons not to cut off unverified users: reported user counts, the Roblox stock reaction, EU and US biometric privacy law, and the roughly 850 million people worldwide without ID.

Question. By the last day of each quarter, has a social or content platform with at least 100 million monthly users required every account to pass proof-of-personhood verification, or banned all AI-generated content?

Definitions.

• Social or content platform: a single online service whose principal functions include social interaction or publishing, sharing, or distributing content to an audience, including video, music, image, and text services. A parent company's combined audience across separate services does not count. A community, channel, country edition, or optional human-only mode within a larger service is not a separate platform unless its own monthly audience independently meets the threshold.
• At least 100 million monthly users: at least 100,000,000 monthly active users, monthly unique users, or an explicitly equivalent monthly audience reported by the platform operator for that service. Registered accounts, downloads, subscriptions, visits, impressions, and summed regional audiences without deduplication do not substitute for monthly users. A directly reported daily active audience of at least 100 million also establishes the monthly minimum. The latest reported qualifying audience observation must end no more than 12 months before the policy takes effect or must cover a period during which the policy is in force. The Observation period convention applies.
• Proof-of-personhood verification: a mandatory check intended to establish that an actual living human controls or is responsible for the account, using a human-bound identity or personhood credential, verified identity document, biometric/liveness check, or comparable person-level attestation. Public disclosure of legal identity is not required, and one-account-per-person uniqueness is not required. Email or telephone confirmation, payment, ordinary CAPTCHAs, self-declared age, and possession of a passkey alone do not count.
• Every account: all new and existing accounts worldwide, including organization, advertiser, developer, and automated accounts where those account types are permitted. An organization or automated account may comply through a verified human responsible for it. Existing accounts must complete verification or lose all logged-in access after a finite, published transition period; merely restricting posting, monetization, or selected features does not count. Logged-out browsing may remain available.
• All AI-generated content: all published or distributed content containing expressive text, imagery, audio, or video generated by a generative AI system, including human-edited AI output. This does not include human-authored material merely processed with spelling correction, noise reduction, conventional image correction, or similar non-generative assistance. The ban must cover every supported content format, accounts of every type, advertising, and platform-supplied content. Private person-to-person communications are outside this definition.

Resolves YES if the platform operator publishes an authoritative policy, terms of service, or official policy announcement establishing either of the following, and the audience threshold is documented:

1. A platform-wide personhood-verification requirement meeting the definitions above has taken effect and its transition period has ended; or
2. A platform-wide prohibition on all AI-generated content has taken effect, with violations subject to removal or account sanctions. Existing AI-generated content must also be prohibited, with any grandfathering or transition period ended.

Either branch suffices. Perfect detection or compliance is not required. An official publication must confirm that the qualifying policy is in force; a future effective date alone does not resolve YES. If audience evidence is published later, resolution waits for that evidence.

Exclusions (resolve NO). Voluntary verification; checks limited to suspicious accounts, particular jurisdictions, age groups, or account classes; rules limited to new accounts; community-level bans; AI labels or filters; demonetization or reduced distribution; bans limited to deceptive, low-quality, political, impersonating, or unlicensed AI content; and bans on AI music or images that leave other supported AI-generated formats permitted. Exceptions permitting labeled AI output, satire, educational examples, or AI-generated advertising prevent the content-ban branch from qualifying. Trials and proposed policies do not count.

Resolving source. The platform operator's published terms, community/content policies, help-center rules, or official policy announcements establish the requirement or ban; operator-authored audience reports, investor disclosures, or regulatory filings establish audience size. These are the exclusive resolving sources, narrower than Public confirmation (A2). Apply Resolution date (A3), the Resolution window, and the Observation period convention. Where multiple documents are necessary, use the publication date of the last necessary qualifying document.