How to Compare AI Companion Report, Block and Support Controls in 2026

How do you create a fair comparison worksheet?

Set the same test cases and evidence fields before opening provider pages.

1. Create columns for report path, block or mute path, support channel, escalation information, account security, privacy, billing and deletion.
2. Use three neutral scenarios: an unwanted reply, a suspected privacy exposure and a billing or account-access problem.
3. For each provider, record the exact help-page title, URL and access date.
4. Mark whether a control is visible before signup, available only after login or missing from the public documentation.
5. Set a stop rule before the test, such as rejecting any service with no reachable safety or support channel for your intended use.

Expected result: you compare the workflow, not just the presence of a shield icon or a broad promise. Do not paste provider policies into another AI service for analysis if your worksheet can be completed with an offline note.

How do you check report controls?

Find the report route, the available reason categories and the confirmation you receive without submitting private material unnecessarily.

1. Search the official help pages for report, safety, moderation, abuse, harmful content and appeal instructions.
2. Check whether reporting covers a generated reply, a public character, a user profile, an image, a voice output or only a technical bug.
3. Record what information a report requests and whether you can describe the issue without uploading intimate messages or identity documents.
4. Use a harmless fictional scenario if the provider offers a safe test path; do not create a false accusation against a real person.
5. Save the confirmation screen or case reference without publishing it, and record what the provider says happens next.

Expected result: you know what the report channel is designed to handle and what evidence it asks for. A report is not a substitute for immediate help, and an AI companion should never be asked to investigate or retaliate against another person.

How do you test block and mute behavior?

Test the scope of the control: what stops, where it stops and what remains visible to you or others.

1. Read the provider's current instructions for block, mute, hide, delete and public sharing because these actions may have different effects.
2. Record whether the control applies to a user, character, conversation, notification, recommendation or shared link.
3. Use only a fictional or low-risk account interaction when a test is appropriate; never provoke a real person to create evidence.
4. Check whether the blocked content remains in your history, exports, device notifications or public pages.
5. Confirm how to undo the action and how to contact support if the control does not behave as documented.

Expected result: you can tell the difference between reducing contact and deleting data. Do not claim that blocking removes every copy, and do not use another account to get around someone's boundary.

How do you compare support and escalation?

Look for a reachable channel, a narrow scope and a record of your request, then verify what support is not able to do.

1. Find the official support route without clicking an unsolicited link in a chat or email.
2. Record whether support handles safety, account access, privacy requests, billing and content reports separately.
3. Check whether you receive a case number, an expected next step or instructions for adding evidence.
4. Ask a narrow pre-sale question about a documented control, not whether the provider can guarantee safety.
5. Record the answer and date, then compare whether the response matches the public policy.

Expected result: you know how to get help for an ordinary product problem and where the boundary of that help lies. For threats, stalking, blackmail or immediate danger, use appropriate local support and emergency channels instead of waiting for a commercial support queue.

What privacy checks belong in a safety-control test?

Test the control with the least data possible and check how reports, support tickets and attachments are handled.

1. Do not upload a full chat, face image, exact location, financial record, identity document or intimate media just to see whether reporting works.
2. Check whether reports and support tickets are retained, shared with service providers or used for moderation or improvement.
3. Ask how to remove an attachment or request deletion if your case includes sensitive content.
4. Use a separate note to record the issue; do not rely on a provider chat's memory to preserve your safety plan.
5. After testing, remove synthetic test content, screenshots and downloaded confirmations from shared or synced devices.

The FTC explains that apps and websites may collect information through account activity, permissions and tracking. Treat a safety report as sensitive data. If a provider requires more personal information than the issue warrants and does not explain why, pause before submitting.

How do you include cost and account closure?

Compare controls at the plan and platform you will actually use, and verify that cancellation does not end your ability to resolve a safety or billing issue.

1. Check whether report, block, support, export and deletion controls are available on the free path or only after payment.
2. Record message, memory, media and voice limits that could affect how much evidence you can review or retain.
3. Identify the billing channel, renewal timing and cancellation route before starting a trial.
4. Save only the billing proof and support reference you need, then remove unnecessary copies.
5. Read account-closure and deletion instructions before paying, including any exceptions for security, legal or billing records.

Our methodology separates privacy, chat, memory, media and value. Affiliate placement does not change a provider's safety controls or editorial caveats.

How should you score the comparison by use case?

Weight the controls that matter to your workflow instead of giving every provider the same generic safety score.

1. For casual text chat, prioritize a clear report path, mute control, privacy policy and account recovery.
2. For public characters or social features, prioritize profile reporting, blocking scope, share-link control and moderation documentation.
3. For voice or image use, add microphone, upload, transcript, media deletion and support questions.
4. For long-term memory or journaling, add retention, export, deletion and human-review questions.
5. Reject any result where the unknowns affect a core risk. Do not let good chat quality compensate for a missing control you need.

Safety controls are useful when they make a problem easier to contain and explain. They are not proof that a service can identify a scammer, prevent every harmful output or provide professional care.

Summary: verify the control, not the promise

Find the report path, test block scope, document support, minimize evidence and reject a service when a needed control is unclear.

A report button can be useful, but the real comparison is the complete workflow from problem to containment, support and deletion. Keep serious evidence private, never retaliate or bypass another person's boundary and use human or local help when the situation is urgent.