What the companion is

The companion is an AI, offered to a client by their practice. It is not a person, a therapist or an emergency service. It does not diagnose, prescribe or treat, and it cannot change a client’s care plan. Those decisions stay with the clinician.

Clear from the start

Before a client’s first conversation, they are shown what the companion is, what is checked and what their practice is told. They confirm this before the companion replies, can read it again at any time in settings, and can stop using the companion whenever they choose.

What is checked

Each message a client sends, and their answers to some questionnaires, are checked automatically by software for signs that they may be at risk. The check is separate from the companion itself, and no one reads conversations as they happen.

When a risk is found

Depending on the level of concern, the client may be shown crisis services they can contact straight away, and their practice may receive an alert with the level of concern and what raised it. Each practice decides how it responds at each level.

Only clinicians at the practice who are permitted to see safety alerts can open one. An alert shows the message that raised it and the few messages around it, not the client’s full conversation.

What does not happen

The companion and Althio do not contact emergency services on a client’s behalf. Conversations are not monitored live, and a practice may not see an alert straight away, especially outside its working hours. Anyone who needs help now should contact their local emergency services.

Privacy

Conversations are encrypted in transit and at rest, and access is limited to the client’s practice. Clients can turn off the companion’s memory of earlier conversations, and can ask their practice about the information it holds about them.

Read about security and privacy

AI safety at Althio