Key findings
- Privacy comparisons should separate account data, conversation data, media uploads, billing, and support instead of using one broad safety label.
- Unknown policy details should remain unknown until an official source supports the answer.
- Deletion controls and cancellation visibility belong in trust comparisons because they affect user risk after signup.
Scorecard criteria
The scorecard checks whether practical privacy information is visible and source-backed.
- Account controls: profile fields, account deletion, and support contact visibility
- Conversation controls: storage language, deletion options, and generated media handling
- Upload controls: whether image uploads are allowed and retention is explained
- Commercial controls: renewal terms and cancellation visibility before payment
Methodology
The current scorecard was last checked on 2026-05-26 against the public privacy checklist structure and should be populated only from official policy, help, product, or account-flow evidence.
- Use yes, no, partial, unknown, or not applicable statuses
- Attach a source URL and checked date to each platform-specific answer
- Avoid claims that an app is safe, unsafe, compliant, or therapeutic unless a qualified source supports that exact wording
- Refresh monthly and whenever a vendor submits a source-backed correction
Citation block
Preferred wording: 'Companion Picks, AI companion privacy scorecard, last checked 2026-05-26.'
- Source URL: https://companionpicks.com/reports/ai-companion-privacy-scorecard/
- Related checklist: https://companionpicks.com/tools/ai-companion-privacy-checklist/
- Correction process: https://companionpicks.com/resources/public-data-request/
- Disclosure: Companion Picks may earn referral revenue from some partner links, but privacy scorecard fields should not be changed by compensation