Skip to content
Talent Connect World

Proposed revision 2026-09-23 · Draft, pending counsel review · Not effective

Proposed legal revision 2026-09-23

This proposed revision describes how automated text, image and sampled video-frame safety checks work, what those results can and cannot prove, and the categories of information processed for safety. It is published for review only. The 2026-08-11 version stays in force until qualified counsel approves a replacement and Talent Connect World LLC records that approval.

Terms of Service

Adds a plain description of automated safety checks applied to uploads and messages, and the fail-closed outcomes they can produce.

Read the controlling Terms of Service (version 2026-08-11)

Proposed: automated safety checks on submitted material

Text, images and sampled frames taken from supported video may be inspected by automated safety systems, including services operated by third-party vendors under contract to Talent Connect World LLC, before or shortly after submission.

An automated result is a safety signal, never proof that content or a person is unlawful. A result that is ambiguous about sexual content, the possible involvement of a minor, or the context of a depiction keeps the item private for trained human review instead of producing a public decision.

Malware detection and a confirmed match against a known-illegal-image hash set always fail closed: the item stays private and quarantined, and it cannot be published, sold or delivered. A provider outage, timeout, unreadable file or unsupported format is recorded as a technical error and also fails closed; it is never recorded as a clean result.

Notice, appeal, accessibility, privacy and data-rights protections in the controlling agreement are unchanged. A human reviewer decides any removal, restriction or account action that follows an automated signal, and lawful preservation of evidence continues where TCW is required or permitted to preserve it.

Proposed: what an automated result means for your account

A pass result means the checks that ran returned no finding for the categories they cover. A review result means the item stays private until a trained reviewer looks at it. An error result means inspection did not complete. None of these three outcomes is a finding that you broke a rule.

  • An automated signal alone is not a breach determination and is not published to other members.
  • Repeated confirmed violations remain subject to the existing enforcement and appeal process.
  • TCW does not charge for a safety review and does not treat a technical error as a violation.

Privacy Notice

Truthfully describes the categories processed for automated safety, the purposes, vendor disclosure, access restriction and the retention framework, without promising controls TCW cannot evidence.

Read the controlling Privacy Notice (version 2026-08-11)

Proposed: information processed for content safety

To run safety checks, TCW processes the file or message you submit, derived values computed from it such as image hashes and sampled video frames, technical file metadata, the safety outcome, the time of the check and an opaque provider reference. Where the material you submit contains personal information, that information is processed as part of the check.

The purposes are limited to detecting malware, detecting known illegal images, detecting material that is prohibited or age-restricted under the Community Rules and Content and Moderation Policy, protecting members and complying with legal obligations.

  • Safety vendors act as service providers or processors for TCW and are permitted to use the material only to return a safety result to TCW.
  • Access to quarantined material inside TCW is restricted to trained, multi-factor-authenticated reviewers, and each access is recorded.
  • TCW stores safety outcomes, digests and provider references. It does not store government identification images or dates of birth from identity checks.
  • Retention follows the retention framework in the controlling Privacy Notice: safety outcomes and audit records are kept for as long as needed for safety, dispute, appeal and legal-obligation purposes, and the underlying material is deleted on the schedule that applies to the post, message or asset it belongs to, unless a lawful hold applies.

Proposed: limits TCW states honestly

TCW does not claim that a safety vendor keeps no logs, retains nothing, or never uses submitted material to improve its own systems beyond what its contract with TCW allows. TCW does not assert a specific legal basis, a signed data-processing agreement or a certified security control for a vendor in this notice unless the corresponding evidence exists and is referenced in TCW's records.

Safety checks may involve processing in countries other than your own, including the United States. Where a transfer mechanism, representative or additional safeguard is required for your country and is not yet in place, the affected feature stays unavailable in that country instead of being run without the safeguard.

Your rights to access, correction, deletion, objection, restriction, portability and complaint under the controlling Privacy Notice apply to safety records, subject to lawful exceptions such as fraud prevention, an active investigation or a legal preservation duty.

Notice at Collection

Adds the safety-processing categories and purposes to the collection notice.

Read the controlling Privacy Notice at Collection (version 2026-08-11)

Proposed: safety category and purpose

When you upload a file or send a message, TCW collects that content, values derived from it such as hashes and sampled frames, technical metadata and the resulting safety outcome.

The purpose is safety, fraud and malware prevention, legal compliance and rule enforcement. This material is not sold, and it is not shared for cross-context behavioural advertising.

  • Recipients are limited to contracted safety vendors acting as service providers and to trained TCW reviewers.
  • Retention follows the framework described in the Privacy Notice.
  • Requests about this category use the same rights process as every other category.

Regional Privacy Disclosures

Confirms regional rights apply to safety processing and that unsupported regions stay closed rather than processed without safeguards.

Read the controlling Regional Privacy Addendum (version 2026-08-11)

Proposed: regional treatment of safety processing

Regional rights described in this document — including access, deletion, correction, objection, restriction, portability, appeal of an automated outcome and complaint to a supervisory authority or attorney general — apply to content-safety processing.

Where a region requires a transfer mechanism, local representative, specific legal basis or additional disclosure that TCW has not put in place and cannot evidence, the affected feature remains unavailable in that region. TCW does not treat a page being reachable as permission to process.

A human review route is available for any safety outcome that materially affects publication, sale or account status.

Cookies and Similar Technologies

Clarifies that safety checks are server-side and not cookie-based tracking.

Read the controlling Cookie and Similar Technologies Notice (version 2026-08-11)

Proposed: safety checks are not cookie-based tracking

Content-safety checks run on TCW servers and with contracted safety vendors after you submit material. They do not rely on advertising cookies, cross-site identifiers or device fingerprinting, and declining optional cookies does not disable them.

Strictly necessary storage may still be used to keep you signed in while an upload is inspected and to show you the resulting status.

Community Rules

Explains what automated states mean to members and preserves appeal rights.

Read the controlling Community Rules (version 2026-08-11)

Proposed: how automated checks affect your post

Automated checks may hold your upload or message while inspection completes. A hold is not an accusation.

An automated result is a safety signal, never proof that content or a person is unlawful. A result that is ambiguous about sexual content, the possible involvement of a minor, or the context of a depiction keeps the item private for trained human review instead of producing a public decision.

You keep the right to be told when something is removed or restricted, the reason in plain language, and how to appeal to a human reviewer.

Content and Moderation Policy

Describes the automated layer, the human review layer, the fail-closed cases and the escalation route while external reporting credentials are pending.

Read the controlling Content and Moderation Policy (version 2026-08-11)

Proposed: automated layer and human layer

TCW operates automated checks for malware, known illegal images and text, image and sampled-video-frame classification, followed by trained human review for anything the automated layer cannot resolve.

An automated result is a safety signal, never proof that content or a person is unlawful. A result that is ambiguous about sexual content, the possible involvement of a minor, or the context of a depiction keeps the item private for trained human review instead of producing a public decision.

Malware detection and a confirmed match against a known-illegal-image hash set always fail closed: the item stays private and quarantined, and it cannot be published, sold or delivered. A provider outage, timeout, unreadable file or unsupported format is recorded as a technical error and also fails closed; it is never recorded as a clean result.

Proposed: suspected child sexual abuse material

Material suspected to depict a minor is immediately quarantined, kept out of public view, and preserved under access controls with every access recorded. Access is limited to trained reviewers and, where required, to counsel and lawful recipients.

A report is submitted only through an approved, registered reporting channel once the operator holds issued credentials for that channel. Suspected material is never sent by email, never attached to an ordinary support message, and no report identifier is ever recorded unless the receiving authority issued it.

Until that registered channel is available to TCW, affected adult features stay closed, the material stays quarantined and preserved, and the matter is escalated through the documented safety and counsel process.

Proposed: notice, appeal and accessibility

Notice, appeal, accessibility, privacy and data-rights protections in the controlling agreement are unchanged. A human reviewer decides any removal, restriction or account action that follows an automated signal, and lawful preservation of evidence continues where TCW is required or permitted to preserve it.

Appeal decisions are made by a person who did not make the original decision where staffing allows, and the outcome is explained in plain language in an accessible format.

Copyright and Intellectual Property

Separates automated safety signals from copyright determinations.

Read the controlling Copyright, Trademark and Rights Policy (version 2026-08-11)

Proposed: automated safety checks are not copyright screening

Automated safety checks look for malware, known illegal images and rule-restricted categories. They do not determine ownership, licensing or fair use, and a pass result is not a statement that you hold the rights to the material.

Copyright complaints, counter-notices and repeat-infringement handling continue to follow the notice-and-appeal process in the controlling document.

Marketplace, Booking and Client Terms

States the effect of a safety hold on listings, orders and delivery.

Read the controlling Marketplace, Booking and Client Terms (version 2026-08-11)

Proposed: safety holds on listings and deliverables

A listing image, deliverable file or message may be held while safety checks run or while a trained reviewer looks at an ambiguous result. A held item cannot be published or delivered.

A hold pauses the related delivery deadline rather than creating an automatic breach for either side. Where a held item cannot lawfully be delivered, the order is cancelled and refunded under the refund terms instead of being silently failed.

Malware detection and a confirmed match against a known-illegal-image hash set always fail closed: the item stays private and quarantined, and it cannot be published, sold or delivered. A provider outage, timeout, unreadable file or unsupported format is recorded as a technical error and also fails closed; it is never recorded as a clean result.

Creator and Professional User Terms

Tells creators what the automated layer checks and what it cannot certify.

Read the controlling Creator and Professional User Terms (version 2026-08-11)

Proposed: what creators should expect from safety checks

Everything you publish or send may be inspected by automated safety systems. A pass result means no finding in the categories checked; it is not clearance that your material is lawful, licensed, correctly labelled or suitable for every audience. You remain responsible for rights, consent, labelling and applicable law for your own work.

An automated result is a safety signal, never proof that content or a person is unlawful. A result that is ambiguous about sexual content, the possible involvement of a minor, or the context of a depiction keeps the item private for trained human review instead of producing a public decision.

You can appeal any restriction to a human reviewer, and you keep access to your own material except where a lawful hold prevents it.

TalentWorld X Adult Content Policy

Records that the automated content-safety gate passed while keeping the zone closed and separating that gate from overall approval.

Read the controlling TalentWorld X Adult Content Policy (version 2026-08-11)

Proposed: the automated content-safety gate is separate from approval

The automated content-safety gate for the restricted zone has passed its production verification. That single gate is not approval of TalentWorld X. The zone stays disabled, isolated, opt-in and limited to verified adults.

General Talent Connect World remains free of explicit adult material. Restricted material never appears in general discovery, search, sitemaps, recommendations, email or public previews, and never receives a public file link.

  • Age and identity verification: verified.
  • Automated content safety: verified.
  • Adult payment processing: not approved.
  • Specialist adult-safety moderation coverage: not contracted.
  • Attorney-approved country and minimum-age matrix: not approved.
  • Verified encrypted consent and recordkeeping vault: not verified.
  • Registered child-safety reporting channel: registration requested, credentials not issued.
  • Completed incident-response drills: not completed.

Proposed: fail-closed behaviour while approvals are outstanding

Any missing approval, credential, consent record, verification, scan result or operator decision denies access or publication. Suspected illegal material is quarantined and preserved under access controls and escalated through the documented safety and counsel process, and no report identifier is created unless a receiving authority issued it.

Global Availability and Country Controls

Reaffirms that no country is enabled for the restricted zone and that reachability is not legality.

Read the controlling Global Service Availability Notice (version 2026-08-11)

Proposed: restricted-zone country availability

No country or state is enabled for TalentWorld X. A country becomes eligible only after qualified counsel approves it in writing for TCW's exact model and an approved payment processor confirms it. Both approvals are required, and either can be withdrawn.

A page being reachable from a location is never a statement that a feature is lawful or permitted there. Where a required safeguard, notice, language, representative or age control is missing for a country, the feature stays unavailable in that country.