Yaziio logoYaziio

AI and Automation Disclosure

Last updated: 4 October 2026

When you message an account that uses Yaziio, we want you to know the following.

Yaziio is in development: the Meta app is not live yet, automatic replies are switched off and every reply goes through human approval before it is sent; it does not yet send replies to any comment or message. What is described below will apply once the service goes live; the relevant rules are implemented in the software and covered by tests. Sending automatic replies is a planned feature.

Automated assistant

Once automatic replies go live, an automated assistant may reply to your comment or message on an account that uses Yaziio. The assistant does not claim to be human.

Automatic-reply note

A single-line note is added to the end of the first message in a conversation: "Bu yanıt otomatik gönderildi. İnsanla görüşmek için 'insan' yaz." (Turkish for "This reply was sent automatically. To talk to a person, write 'insan'."). The account owner can switch the note off in settings; it is on by default.

The "are you a bot?" question

The assistant does not greet or introduce itself and goes straight to the point. But if you ask "are you a bot?" or "is a real person writing?", it honestly says it is an automated assistant.

Reaching a human

At any time you can write "insan" (Turkish for human; also "yetkili" or "temsilci") to ask for a person; the conversation is handed to the account owner and the assistant stays silent in that conversation. Conversations the assistant is unsure about, complaints and sensitive topics are also handed to the account owner. For now every reply is in any case approved by a person before it is sent.

Stopping messages: DUR

If you do not want to receive messages from Yaziio, just write DUR (Turkish for "stop") in the Instagram DM; STOP, İPTAL, "bana yazma" and "abonelikten çık" work the same way. The whole message must be the command; sentences such as "I want to cancel my registration" count as a support topic, not as a stop command. Once the service goes live it works like this (the compliance desk is implemented in the software and covered by tests; there is no live message flow yet):

Limits (design principles)

Your data and AI

Once reply generation is live, your message may be sent to Anthropic's AI model (Claude) to draft a reply and determine the topic of a message; because no live user messages flow today, no real user message is sent to an AI model. Yaziio does not use your messages to train its own models. The provider's data handling is governed by its own terms. Details: Privacy Policy.