Character.AI's Terms of Service document appears to have been corrupted or improperly captured during the June 16, 2026 update. The 'AFTER' version contains HTML metadata and Next.js framework code rather than readable Terms of Service language. The substantive terms content is not visible in the provided diff, making it impossible to identify what specific operational changes were made to the agreement.
The provided change summary does not contain readable Terms of Service language. The 'AFTER' version consists of HTML framework code and metadata rather than policy text. Without access to the actual updated terms, no operational impact on consumers can be assessed.
The document capture is corrupted and contains no readable policy language. A proper assessment requires retrieving the actual Terms of Service text.
This change record describes what was added, removed, or modified in the document. Analysis reflects what the updated agreement states or permits. It does not constitute a legal determination about enforceability. Applicability may vary by jurisdiction. Methodology
A Terms of Service update was detected on June 16, 2026, but the document capture is malformed. The diff contains HTML and Next.js framework code rather than substantive policy language. Compliance teams cannot evaluate the change until the actual policy text is properly captured. No regulatory exposure or obligation changes can be determined from this corrupted data.
Full compliance analysis
Obligation analysis, escalation trigger, board language, and recommended action.
Monitor: regulatory citations + obligations. Compliance: full compliance memo.
ConductAtlas provides verified policy intelligence sourced directly from platform documents. All analysis is intended to support, not replace, legal and compliance review. Record CA-C-002990.
New provision explicitly clarifies user ownership of AI-generated outputs while still granting Character.AI a perpetual sublicensable license, addressing a gap between user-submitted and AI-generated content rights.
Privacy-related disclosure provision removed from top-level terms structure, though content moderation language in current version may partially address this through law enforcement reporting clause.
Standalone AI content disclaimer removed, likely consolidated into the new 'AI-Generated Content License' provision which now addresses both ownership and licensing rather than just disclaimers.
Standalone price increase notice removed as a separate provision, consolidated into the 'Subscription Non-Refund Policy' section in current version.
Previous version had empty excerpt; current version now provides full arbitration clause text with explicit exception for small claims.
Previous version had empty excerpt; current version now explicitly details the perpetual license grant for user-submitted content.
Previous version had empty excerpt; current version now provides detailed liability limitation language with damages exclusions.
Previous version had empty excerpt; current version now explicitly details broad indemnification scope including attorneys' fees and injuries.
Previous version had empty excerpt; current version now explicitly specifies age thresholds including EEA/UK GDPR compliance (16 years for EEA/UK residents).
Previous version had empty excerpt; current version now adds legal carve-out ('except where required by law') and consolidates price increase notice into subscription policy.
Previous version had empty excerpt; current version now provides detailed enforcement procedures including DMCA compliance language.
Cross-platform context
See how other platforms handle similar provisions across the ConductAtlas archive.
See the full side-by-side comparison of every sentence added, removed, and modified.
🔒 Full diff — MonitorCharacter.AI removed language from its Safety Center footer on June 16, 2026. The previous version included a link to 'About …
Character.AI's privacy policy document was updated on June 16, 2026, with 200 sentences added and 99 sentences modified. However, the …
Get alerted when this policy changes again — including what changed and why it matters.
Prefer a weekly summary instead?
Get the biggest policy changes across 320+ platforms every Sunday.