Character.AI Privacy Policy
- Document
- undated document
- Event
- no single event
- Retrieved
- 16 September 2026
The design
Character.AI, operated by Character Technologies, Inc., publishes a Privacy Policy that, as retrieved on 16 September 2026, covers the websites, apps, and offline services that link to it. Section 7, “Data Retention,” says the company keeps information “for the time necessary for the purposes for which it is processed,” language that sets a standard rather than a number of days. A separate clause carves out an exception: if a Character a user created has been made available to other people, Character.AI “reserve[s] the right to preserve that Character's characteristics and to keep that Character active on the Services,” even after the creator deletes their account and data.
What the evidence says
The policy itself supplies the retention standard; it does not point to an external audit of how long any specific message actually persists. The company's Regional Privacy Disclosures add that people in the European Economic Area or the United Kingdom have a right to opt out of having their data used to “train and improve” Character.AI's models, and that this is separate from the one-time collection used for the “initial development” of its earlier models, which the company says it “no longer carry[ies] out.” Both statements come from Character.AI itself.
What it asks of people
A user who deletes their profile can expect most personal information to go, per the policy's general deletion provision, but a Character they built and shared with others can outlive that deletion by design. The policy asks users creating a shareable Character to avoid embedding identifying details in it for exactly this reason. Deletion requests can also be refused where the company says it needs the data “to provide the Services” or to meet legal obligations, a standard exception across many of these policies rather than one unique to Character.AI.
Privacy and safeguards
The policy discloses a broad internal use for retained data: “Analyze, maintain, improve, modify, customize, and measure the Services, including to train our artificial intelligence/machine learning models.” It does not state a retention ceiling for ordinary conversation data, and it does not describe any outside body that checks compliance with the purpose-based standard it sets for itself.
- What purpose would justify keeping a specific conversation on file five years from now?
- Can a user confirm that a shared Character no longer contains their personal details after they delete their account?
- Does the general EEA/UK opt-out right described in the Regional Disclosures reach every use of conversation data?
Character.AI's retention language sets a purpose-based test rather than a fixed clock, and its own carve-out for popular Characters is one of the more specific disclosures in this batch of policies — useful precisely because it names a case where deleting an account does not delete everything tied to it.
Sources & reading trail
States the purpose-based data retention standard, the Popular Characters exception that can survive account deletion, and the listed uses of information including AI model training, as the policy read on 16 September 2026.
Source published: Not established · Retrieved: 16 September 2026
States that Character.AI trained its proprietary models primarily on publicly available data, that it no longer carries out that initial-training collection, and that EEA/UK users have a right to opt out of ongoing model-training use.
Source published: Not established · Retrieved: 16 September 2026
Product documents, regulator records and studies establish the entry; the design reading is AI Companions editorial analysis. This retrospective draft does not imply the site published on the event date.