Is Character.ai safe?

A lot of groups have raced to exploit the hype and legitimate technical advantages of generative AI. One of the more interesting experiments in that regard is Character.ai, which lets you use and create chatbots that simulate fictional characters, real celebrities, or helpful assistants. You can chat with everyone from anime heroes or Friedrich Nietzsche to a programming assistant or a debate champion. But is Character.ai safe to use in terms of privacy, security, or potentially offensive content?

Is Character.ai safe?

Generally speaking, yes. While you do have to create an account to take full advantage of it, the service is free for now, so there’s no direct monetization of your activities from what we can tell. Its developer, Character Technologies, does mention allowing ads and trackers in its privacy policy, but it doesn’t seem to be any worse than most web companies, and there aren’t any ads in place yet. You can delete your account if necessary.

The site also uses SSL encryption, and has a moderation team that can review public material for potential terms-of-service violations (more on that below). If you don’t like what’s going on, whether in chats or community posts, there are ways of reporting it.

Something worth considering if you’re a parent is that Character.ai explicitly disallows users under 13, or under 16 if they’re in Europe. In fact kids under 13 are supposed to be banned, since taking their personal info would violate the US Children’s Online Privacy Protection Act, but that’s contingent on Character Technologies catching someone being dishonest about their identity — a tough thing to do. It’s like Activision trying to catch a 12-year-old playing Call of Duty.

Are Character.ai chats private?

That depends on the privacy settings you choose. If something you share is set to public, moderators may decide to review it, and of course everyone on the site (or sometimes even outside the site) will be able to see it. You can keep chats and characters fully private if you want to however, and hide recent chats by clicking Edit in the Chats tab. Down the road, Character Technologies is promising the option to delete chats entirely.

Does Character.ai allow NSFW content?

NSFW, if you’re not familiar, means Not Safe for Work. At the moment NSFW content is completely off-limits, and Character Technologies states that pornography will never be allowed. Trying to create that sort of material is probably a quick ticket to getting banned, assuming it makes it past any filters in place.

Swearing could be allowed sometime in the future, but only as an option. Indeed the site’s AI is almost prudish at the moment — if you try to create a villain chatbot, they’re still be going to be relatively polite, even if they should be losing their temper and threatening to crush everyone you love. That quirk should be fixed sometime in the future.