About this simulation

What it does not do

It does not connect to real phone calls. No AI will answer your phone. It cannot block incoming calls either.
It does not judge whether a call is a scam. There are only six scenes, all fiction, and real calls do not go the way they do here.
Talking the way shown here does not make you safe. We do not promise that you will avoid harm. If you are in trouble, please consult a public office such as the police or a consumer affairs center.

This is a simulation for people who hesitate over whether to answer calls from unknown numbers, and for their families, to look at how such calls can go while things are calm.
When you press "Let the AI answer", you can read and hear how an AI might reply in your place: asking for the caller's name and reason, not answering about money or PINs, and ending the call. You can also choose your own reply or hang up at any time.
This is new. There are no user comments or numbers yet that show it helps.

The "AI" on this page only shows lines that were written in advance. It is not an AI thinking and answering each time you press a button. Nor does it show that an AI would reply this way on a real call.
The caller's lines, the AI's replies, your replies, the one-line comments and the example notes are fiction written by AI. They have not been checked by experts such as the police, lawyers or consumer affairs advisers. The English text was also written by AI and has not been checked by a native English speaker.
The types behind the scenes were chosen from what a public agency in Japan (the National Police Agency) has published. The sources are listed in each review and in "Checks before you answer". No real company, government office, bank or person's name, and no phone number, appears in the scenes. The English versions of the agencies' text were written by AI. This page was not made, checked or approved by the National Police Agency or any other public agency.
Some scenes were written as harmless calls. Not every call from an unknown number is a scam.
The voices, for both the caller and the AI, are synthesized. They are not the voices of real people. No person has listened to check that the readings are correct.
It has not been checked on an actual smartphone.

When you open the page, one call scene starts, picked at random. Choose with the buttons: "Let the AI answer", your own reply, or "Hang up". When the call ends, you see a review and choices for the next scene. Nothing moves on until you choose.
The caller's lines and the AI's replies appear as text and audio. Your own replies are text only. Audio plays when you open the page and when you press a button (some browsers do not play it on opening). You can turn it off with the "Sound" button at the top. You can use the page with text only if there is no sound.
No sign-up. Free.

No cookies are used. Nothing is stored in your browser. The server does not create a record for you as a person (no ID, name, email address or phone number).
The server keeps daily totals of how many times a scene was opened, "Let the AI answer" was pressed, an own reply was chosen, a call was hung up, a scene was followed to the end, and "Checks before you answer" was opened, counted by screen language and scene. They cannot be linked to any person. In addition, ordinary access logs (date and time, IP address, browser information and the URL of the page) are kept, as with our other services. In this simulation, the scene and the sequence of what you chose are part of the URL, so they also appear in those logs.