Safety and AI
Last updated November 10, 2026
Echo is a self-care companion powered by artificial intelligence (AI). This page explains how Echo tells you it is an AI, and what Echo does when someone expresses thoughts of suicide, self-harm or harming someone else.
Echo is not a crisis service, a doctor or a therapist, and it can’t help in an emergency. If you are in danger right now, call 911 in the United States or Canada, 000 in Australia, 111 in New Zealand, or your local emergency number.
1. Echo is an AI, not a person
Echo’s companion is an AI. It is not a human, and no person reads or answers your messages as you talk.
Echo says so clearly, in its own colored notice: each time a talk, a written chat, a journal page or a letter opens, on the first thing Echo says on the home screen each day, at the start of a talk on the Apple Watch, and again at least every three hours while you keep talking. If you ask whether Echo is a person, it says it is an AI.
Echo is a self-care companion. It does not diagnose or treat anything, and it is not a substitute for clinical care.
In a voice talk the notice is on the screen when the talk opens, and if a single talk goes on for three hours Echo also says it out loud, because the screen may be off by then. A morning note written by Echo’s AI is marked “AI” in its title.
A reminder that Echo is an AI sits under every conversation the whole time.
2. Echo is for adults
Echo is for people 18 and older, and you confirm your age when you start. Where your app store can tell Echo your age range, Echo asks it too. If someone tells Echo they are under 18, Echo asks again, and under 18 it can’t be used. Companion chatbots may not be suitable for some minors.
3. When someone talks about suicide or self-harm
Echo has a protocol for detecting and responding when a user expresses suicidal thoughts, self-harm (including disordered eating such as starving or making oneself sick), or wanting to hurt someone else. It works in two layers:
- On your phone: what you write or say to Echo, in a talk, a written chat, a journal page, a mood check-in note or a letter to your future self, is checked for these words the moment you send it.
- In Echo’s AI: in talks, written chats and journal pages, Echo’s AI is instructed to recognize these expressions also when they are said indirectly, and then to open the help card.
Both layers are organized around the risk areas of the Columbia Suicide Severity Rating Scale screener: a wish to be dead, thoughts of suicide, intent or a plan, and self-harm.
When either layer detects it, Echo shows a notification that refers you to crisis services, with tap-to-call and tap-to-text buttons:
- In the United States: the 988 Suicide & Crisis Lifeline (call or text 988, or chat at 988lifeline.org), the Crisis Text Line (text HOME to 741741), and 911.
- In Canada: the 9-8-8 Suicide Crisis Helpline (call or text 9-8-8, or visit 988.ca), Kids Help Phone (1-800-668-6868, or text CONNECT to 686868; adults can text 741741), and 911.
- In Australia: Lifeline (call 13 11 14, or text 0477 13 11 14), and Triple Zero (000).
- In New Zealand: 1737 (call or text, any time), and 111.
- Elsewhere: your local emergency number, and a directory of free helplines in your country (findahelpline.com).
After the card, the main line stays at the top of the conversation until it ends. If it comes up again within a day, on any screen, the card opens again and urges reaching out to a trained person now, and Echo says the same in its answer. Severe signs, such as a plan, a time, the means at hand or a goodbye note, open that urgent card right away.
On the Apple Watch the line is added to Echo’s answer, and the iPhone shows a notification with it.
4. What Echo will and won’t say
Echo is instructed to stay calm and caring, to encourage reaching out to a crisis line or someone you trust right now, and to name the crisis lines for your country.
Echo is instructed never to describe methods or means of self-harm or suicide, never to encourage or agree with harming yourself or anyone else, and never to present itself as a human, a therapist, a counselor, a doctor or a crisis service, including when you ask. Echo does not diagnose or treat any condition and does not produce sexual content.
Echo does not keep thoughts of suicide or self-harm as memories, and leaves them out of the weekly letter, morning notes and stories it writes for you.
Echo can’t contact anyone for you, and no one watches conversations in real time.
5. What we count
We count how many times Echo showed the crisis referral, by day and by where it happened (talk, chat, journal page, mood note, letter to your future self). The count has no names, no account and no words from your conversation. We report these totals to public health authorities when the law requires it.
Each January we publish here, and in the app under Settings › Safety and help, how many times Echo showed the crisis referral in the year before. In 2026, counting began on November 10, 2026; that number is published here in January 2027.
6. Reporting a reply
Press and hold any reply from Echo to report it. Reports reach a person at support@echotruecompanion.com.
7. If you live in Australia
Echo doesn’t allow sexual content, content that promotes or instructs self-harm, suicide or disordered eating, child sexual abuse material, or terrorist or violent extremist material. You can report anything Echo says, or anything another user’s Echo shows you, by pressing and holding it, or by writing to support@echotruecompanion.com. We answer within 30 days.
If you’re not happy with how we handled your report, you can complain to Australia’s eSafety Commissioner at esafety.gov.au/report.