Found in 145 of 352 platforms tracked (41% adoption) · 376 provisions
Any input ActiveCampaign provides on Marketing Content does not shift responsibility to ActiveCampaign; the user remains fully and solely liable for that content.
Automatic scanning of Cloud Content means all user-stored content is subject to programmatic review for illegal or abusive material, regardless of user consent.
Ancestry's removal authority is triggered by its own belief that content violates the Terms, meaning content can be removed without a formal finding of violation.
Users may lose the ability to access or re-download content they have purchased if Apple removes it from the Services.
Asana has committed to both removal and mandatory reporting to law enforcement upon awareness, establishing a clear enforcement obligation for this category of content.
Users' chat content is subject to filtering based on Character.AI's Terms and Community Guidelines, which can result in content being withheld.
ElevenLabs' reservation of a right to moderate all inputs and outputs means no user content on the platform is exempt from potential review or restriction.
ElevenLabs' reservation of moderation rights over all Input and Output means no user content or generated content is categorically outside its review.
Active monitoring using multiple methods means users' activity on the platform is subject to ongoing review for policy compliance, not merely reactive investigation.
These prohibitions and the stated removal policy are aimed at preserving content integrity; content perceived as incentivized is subject to removal at Glassdoor's discretion based on its belief.
The clause establishes that safety policy details are located in a separate document, meaning readers must consult that document to understand the safety framework governing Gemini 3.1 Pro.
The clause establishes that Gemini 3.1 Pro cleared a defined, expert-developed threshold before launch in the child safety domain, indicating a formal gate was in place and passed.
This clause defines the conditions under which Google's content delivery obligation terminates and establishes the exclusive remedies available when removal occurs. The refund-as-sole-remedy limitati…
The presence of the tag triggers an automatic public-facing display change on the model page, meaning adding or omitting this tag has direct and visible consequences for how the model is presented to…
Users have no guaranteed right to keep their content visible or stored on the platform; Ideogram may act on content without any prior warning or stated justification.
The clause simultaneously permits monitoring of game chat and expressly disclaims any obligation to monitor, meaning users cannot rely on monitoring as a safety guarantee while also lacking privacy i…
The claim establishes a universal safety training baseline across the entire GPT-5 model family, meaning no GPT-5 model is deployed without this layer.
Users have no contractual assurance that their content will remain on the platform; Public.com can act against any content without being required to justify its decision.
The clause establishes the scope of regulatory protections available to different account types. SIPC insurance provides coverage up to $500,000 per customer in the event of broker insolvency, but th…
This establishes that user-submitted content and generated outputs are actively monitored by both machine and human reviewers, meaning no submission is purely private from internal review.
Prioritization indicates that child safety reports receive expedited attention relative to other report types, which affects how quickly such concerns are acted upon.
Harmful content is intercepted automatically before reaching the end-user, providing a system-level safeguard conditional on the content being identified as harmful.
Real-time scanning of generated content for a defined set of harmful categories means potentially harmful outputs are assessed before or as they are delivered, providing a technical safeguard.
Users have no advance notice right and no required justification before Spotify removes or disables their User Content.
The clause protects real-life actors whose likenesses are used as Stock Avatars from being falsely associated with opinions or endorsements they did not actually make.
The trigger is 'reason to believe', not confirmed breach or actual harm, meaning TikTok Ads can act on suspicion alone, including potential future harm.
Because Twitch can act at any time and without notice, users have no guaranteed advance warning before their content is removed or moderated.
This clause gives Yelp unrestricted authority over User Content with no obligation to notify users or justify its moderation decisions.
Users' private messages, including emails sent to eBay alias addresses, are subject to automated filtering and potential human review, limiting message confidentiality.
Cohere's acknowledgment that toxicity remains possible despite safeguards is consequential for users and developers who may encounter harmful outputs during extended interactions.
The existence of active safety features means certain content will be automatically blocked within the Services.
The terms incorporate the Usage Policies by reference and prohibit using outputs to train competing AI models, which has implications for developers and researchers working in the AI space.
Characterizing filtering as part of teaching the AI 'right from wrong' establishes that harmful content controls are embedded in the model's foundational training, not only applied at the surface lev…
OpenAI does not proactively moderate private GPTs; moderation is contingent on customer action, meaning problematic private GPTs may go unreviewed without a customer-initiated report.
A commitment to rigorous bias evaluation signals that content outputs are actively reviewed against a standard of stereotype avoidance, not merely passively filtered.
The authorization establishes the operational framework under which Robinhood Financial engages in principal trading activity rather than acting solely as an agent or broker. This structure directly …
The claim establishes that enforcement is not purely human-reviewed, meaning automated systems play a role in flagging content.
This provision allocates content moderation authority to OpenAI through discretionary determination standards. The clause establishes OpenAI's operational mechanism for enforcing usage compliance and…
The clause establishes the scope and limits of investor protection available through SIPC membership, defining the maximum recovery available for account holders and identifying categories of loss th…
The clause establishes Robinhood's operational rights to process user-generated content across multiple platforms and formats without additional compensation or exclusive restrictions. This supports …
Monitor emails you the same day a platform you choose changes these clauses.
A content moderation clause is a provision in a platform's terms of service or privacy policy governing content moderation-related rights, obligations, or restrictions.
ConductAtlas tracks 145 platforms with content moderation clauses - roughly 41% of platforms in the archive. 118 are classified as high severity.
Severity reflects the magnitude of rights waived, availability of opt-out, breadth of users affected, financial or legal exposure created, and the degree of discretion retained by the platform.