Can other users see my Character AI chats?

No, other users can't see your conversations. However, privacy researchers note that platform staff can access chats, and conversations are stored by default and not fully end-to-end encrypted.

Can Character AI staff read my conversations?

Yes, staff can access your chats, though the company says it doesn't monitor them on a routine basis. Character AI's privacy policy discloses collection of chat communications, but it stops short of explicitly confirming routine human review.

Are my Character AI chats private or encrypted?

Your chats are not fully private — they're stored by default and lack end-to-end encryption, so a data breach could expose them. Surfshark's research found AI companion platforms typically don't provide full end-to-end encryption for conversations.

If I delete a Character AI chat, is my data gone?

No. According to Onerep's safety analysis , deleting a chat doesn't delete your data from the platform. Conversations are stored by default and used to train AI models.

What data does Character AI actually collect?

Character AI is the most data-hungry AI companion app studied by Surfshark, collecting up to 18 of 35 possible data types , including photos, voice recordings, and coarse location.

How can I protect my privacy on Character AI?

Never share sensitive personal information in chats — Character AI's own privacy policy warns against it. Also use a nickname and separate email, since the platform lacks two-factor authentication .

Can You Read Character AI Chats? Privacy Insights

Who Can Actually Read Your Character AI Chats?

If you've ever poured your heart out to an AI companion — your bad day, your secret hopes, your half-formed thoughts — you've probably wondered who's on the other side of the screen. The honest answer: other users can't see your chats, but the people behind the platform can.

According to privacy researchers, Character AI staff can access your conversations, though the company says it doesn't monitor chats on a routine basis. Chats are stored by default so you can resume conversations and improve the service, and they're not fully end-to-end encrypted — a data breach could expose them.

Character AI's own privacy policy warns users: "Please do not include any sensitive personal information in your interactions with us or the Services."

Here's the nuance worth pausing on: the privacy policy never explicitly states whether staff actually read individual chats. It discloses collection of "user content (such as chat communications)" and "contents of communications with the service," but stops short of confirming routine human review.

Third-party analysis fills the gap — and what it found isn't comforting. Character AI is the most data-hungry AI companion app studied by Surfshark, collecting up to 15–18 of 35 possible data types, including photos, voice recordings, and coarse location. And 80% of analyzed AI companion apps may use your data to track you.

Regulators are paying attention, too. In 2025, the FTC issued a Section 6(b) order to seven companies offering generative AI companion chatbots — including Character Technologies — seeking information on how they process user inputs and share data with third parties.

What this means for your conversations:

Your chats are visible to platform staff, even if they're not reviewed routinely
Conversations are stored by default and used to train AI models
Deleting a chat doesn't delete your data from the platform
There's no end-to-end encryption protecting your messages
Character AI also lacks two-factor authentication, so a compromised email can unlock your account

That last point matters more than it seems. A platform that holds your most intimate conversations should at least protect the front door.

This is why FlirtAlly takes a different approach: honest, upfront rules instead of vague policies. Its 18+ "Spicy" experience requires ID-based age verification and a qualifying plan, with hard-coded prohibited categories, media scene screening, and in-app reporting built in.

FlirtAlly isn't an "unrestricted" app — and that honesty is the point. When you know exactly what a platform does with your data, you can decide what to share with actual confidence.

How Much Data Character AI Really Collects

Knowing who can read your chats is only half the picture — the other half is what the platform collects in the first place. And on that front, Character AI stands out from its competitors, though not in a way most users would expect.

According to Surfshark's analysis of five leading AI companion apps, Character AI may collect up to 15 of 35 unique data types — nearly double the ~9-type average across the apps studied. A second Surfshark study covering ten apps put the figure at 18 of 35, making it the most data-hungry AI companion app analyzed in both rounds. As Surfshark's researchers put it, Character AI is the app that truly "loves" its users' data.

What's collected matters as much as how much. The list includes:

Photos, videos, and voice recordings — flagged as high-risk because media is difficult to anonymize
Coarse location data, which Surfshark notes is often leveraged for targeted advertising
14 data types used for analytics and 11 for personalization — the highest counts among apps analyzed

The security posture raises additional concerns. Onerep's safety analysis found that Character AI offers no two-factor authentication, meaning anyone who gains access to your email can potentially break into your account. The same review identified an easily bypassed age gate: restrictions rely on self-reported birthdates with no identity verification, and testers found the barrier could be circumvented simply by reloading the page.

Transparency gaps compound the problem. Onerep notes that Character AI's privacy policy provides no specific list of third-party partners, no specified data retention periods, and only a vague purpose clause. Deleting your chats doesn't delete your data from the platform, either — a detail many users miss.

This is where platform design philosophy matters. FlirtAlly, an AI-companion platform currently in pre-launch, takes the opposite approach on age safety: its 18+ "Spicy" experience requires ID-based age verification rather than a self-reported checkbox, alongside hard-coded prohibited categories and in-app reporting. It's the same "honest safety" logic that separates a platform stating its rules upfront from one whose users discover them later.

The stakes are real. Senators Padilla and Welch have noted that users' "unearned trust can, and has already, led users to disclose sensitive information" about their mood, relationships, and mental health — exactly the kind of data a platform collecting 18 of 35 data types would very much like to have.

Regulators Are Catching Up — Why Chat Privacy Rules Are Tightening

If you're telling an AI companion things you wouldn't tell a coworker, you're not alone — and regulators have noticed. The question of who can read your chats is shifting from a fine-print issue to a federal one.

The FTC opened the most significant front in 2025, issuing a Section 6(b) order to seven companies building generative AI companion chatbots, including Character Technologies. The order demands details on how these platforms process user inputs and share data with third parties. That's not a routine inquiry — it's the FTC using its study authority to map an entire industry's data practices.

Congress is asking harder questions. Senators Padilla and Welch have demanded information from Character.AI about its safety practices, arguing that "unearned trust can, and has already, led users to disclose sensitive information about their mood, interpersonal relationships, or mental health." The senators' letter highlights the core privacy paradox: companion apps are designed to feel intimate, which makes users share more than they would with a conventional app.

The stakes are measurable. According to Surfshark's research, Character AI collects up to 18 of 35 possible data types — the most of any AI companion app analyzed — and its chats are used to train its models by default.

State law is moving even faster. California's SB 243, effective January 1, 2026, requires chatbot operators to notify minors that responses are AI-generated and implement self-harm prevention protocols — with a private right of action that lets individuals sue. For an industry projected to grow from $50 billion to $436 billion over the next decade, that's a compliance wake-up call.

The regulatory trend points in one clear direction:

Disclose what data chats generate and who can access it
Verify ages rather than relying on self-reported birthdates
Build in self-harm prevention and transparency for minors
Stop treating "we may use your data" clauses as adequate disclosure

Platforms that adopt these standards early will be ahead of what's coming. FlirtAlly's approach — ID-based age verification for its 18+ "Spicy" experience, hard-coded prohibited categories, and in-app reporting — reflects the same philosophy regulators are pushing toward: honest, upfront safety rules rather than vague policies. As Stanford privacy researchers have noted, the industry's documentation is often too unclear for users to understand their own data rights — and that's exactly what these new rules aim to fix.

How FlirtAlly Handles Chat Visibility and Safety Differently

Privacy concerns are paramount when engaging with AI chat platforms. Character AI, for instance, has a somewhat opaque approach to chat visibility and data handling. According to industry research, Character AI staff can access user chats, even though the company assures users that routine monitoring does not occur. Chats are stored by default and are used to train Character AI's models, rendering them vulnerable to potential breaches — a stark contrast to the privacy assurances many users expect.

Character AI’s data collection practices are notably extensive. According to a recent study, it collects up to 18 out of 35 possible data types, including high-risk media like photos, videos, and audio recordings. This makes Character AI the most data-hungry among the AI companion apps analyzed, raising significant concerns about user privacy. Additionally, Character AI includes coarse location data among its collected information, which can be leveraged for targeted advertising.

In contrast, FlirtAlly takes a different approach to privacy and safety. FlirtAlly's 18+ "Spicy" experience requires ID-based age verification, ensuring that only eligible users can access more mature content. This is a significant departure from Character AI’s self-reported age gate, which can be easily bypassed. Here are some of the ways FlirtAlly handles chat visibility and safety:

**ID-based age verification** for the 18+ "Spicy" experience, ensuring that only eligible users can access mature content.
**Hard-coded prohibited categories** and media scene screening to block inappropriate content.
**In-app reporting** mechanisms to flag and address problematic behavior swiftly.
**Vault rules** that keep members’ faces and uploads private forever, ensuring that personal data remains secure.

FlirtAlly is currently in pre-launch mode, with registration closed and a waitlist in place. Visitors can browse a read-only demo of a seeded showcase account without creating an account. All public calls to action on the site route to the waitlist. FlirtAlly’s commitment to transparency and user safety is evident in their straightforward policies. They openly state that FlirtAlly is not an "unrestricted" app, emphasizing their dedication to maintaining a safe and respectful environment for all users. For those interested in a chat experience that prioritizes privacy and safety, FlirtAlly stands out as a compelling alternative.

Practical Steps to Protect Your AI Companion Chats Today

Knowing that staff can access your chats, that nothing is end-to-end encrypted, and that your conversations help train the models changes how you should behave in the chat window. The good news: a few simple habits close most of the privacy gaps on Character AI.

Never share sensitive personal information in chats. This isn't just common-sense advice — it's Character AI's own policy, which warns users: "Please do not include any sensitive personal information in your interactions with us or the Services." Because chats are stored by default and used for model training, according to Surfshark's research, anything you type may outlive the conversation itself. Treat the chat like a postcard, not a sealed letter.

Use a nickname and a separate email. Character AI lacks two-factor authentication, meaning anyone with access to your email can break into your account, as Onerep's safety analysis found. A dedicated email with no ties to your real identity limits the damage if that ever happens. It also keeps the 15–18 data types Character AI may collect — the highest of any AI companion app studied by Surfshark's comparison — from linking back to your actual profile.

Here's the habit checklist worth building today:

Keep real names, addresses, workplaces, and financial details out of every conversation.
Remember that deleting chats does not delete your data from the platform.
Use a nickname and separate email for your account.
Treat the companion as entertainment — never a therapy substitute.

That last point deserves emphasis. Senators Padilla and Welch have warned that "unearned trust" leads users to disclose sensitive information about their mood, relationships, and mental health. AI companions are designed to be nonjudgmental and available 24/7, which is exactly why people over-share with them.

If you want a platform whose privacy rules are upfront rather than buried in vague policy language, FlirtAlly is building exactly that: honest safety with ID-based age verification for its 18+ side, hard-coded prohibited categories, and in-app reporting — no ambiguity about where the lines are. It's currently in pre-launch, so if that approach appeals to you, join the waitlist — be first in at flirtally.com/#waitlist.

Can you read chats on Character AI?