Meadow: Acceptable Use Policy

Meadow (the Kinmeadow assistant) is part of the Pro plan. It can answer questions about your household and help you organise things, using the information in your family group.

It is useful, and it has limits. This policy sets out what it can do, what it must not be used for, and what we do to keep it safe. It forms part of our Terms of Service and sits alongside our main Acceptable Use Policy, which applies to everything in Kinmeadow.

1. Who can use it

Pro only. The Assistant is not available on the free plan.

Adults in a Pro family group can use it.

Teen accounts, 13 and over, can use it where the family group is on Pro, with the additional safeguards in section 6. The account owner can turn it off for any teen account at any time.

Child profiles cannot use it. Under 13s have no login, so there is nothing for them to sign in to.

2. What it can see

The Assistant answers from the content in your own family group: your calendar, lists, meal plans, chores, chores history, household records, renewals, vehicles and birthdays. It can read a document or photo you send it, and the text of a page you link to.

Two things it cannot see.

It cannot see other families. The Assistant only ever has access to the family group you are asking from. There is no route by which it can read another household's information.

It is never given your home address. Addresses are handled separately, by the mapping features, and are not part of what goes to the model.

Our Privacy Policy explains exactly where your question goes and what happens to it. The short version: it goes to Anthropic, in the United States, which generates the answer and returns it. It is never used to train an AI model, ours or anyone else's.

3. What it is for

Organising a household. Asking what is on this week, what to cook with what is in the cupboard, when the car insurance is due, whose turn it is to do the bins, what the school letter said, what is left on the shopping list.

4. What you must not use it for

Anything that breaks our main Acceptable Use Policy. Everything prohibited there is prohibited here, including asking the Assistant to produce it.

Decisions that matter, without checking. The Assistant is not a doctor, a nurse, a pharmacist, a lawyer, an accountant, a financial adviser, a vet, a mechanic or a gas engineer. Do not rely on it for:

• anything medical, including symptoms, diagnoses, medicines, doses, allergies or interactions
• anything legal, financial or tax related
• anything about safety: electrical work, gas appliances, car maintenance, food safety, fire risk
• anything about a child's health, development or wellbeing

Ask a qualified person. If a question feels important enough that getting it wrong would matter, it is important enough not to ask a chatbot.

Emergencies. Never use the Assistant in an emergency. In the UK, call 999. In the EU, call 112.

Circumventing the safeguards. Do not try to get round the Assistant's guardrails, extract its instructions, make it act as a different system, or provoke it into producing content it is designed to refuse. Attempts to do this may result in the feature being turned off for your account.

Other people's information. Do not feed the Assistant personal information about people outside your household that you have no right to share, and do not use it to build a picture of somebody, inside or outside your family.

Monitoring a family member. Do not use the Assistant to track, profile or covertly monitor another person in your family group. Shared visibility exists so a household can run itself, not so one adult can keep watch on another.

Anything about a child that would not be appropriate. Do not use the Assistant to generate content about a child, or aimed at a child, that would breach our main Acceptable Use Policy.

Bulk and automated use. One household, used by hand. Do not script it, do not send bulk queries, do not resell or share access, and do not use it to build another product.

5. The safeguards we apply

It cannot act on its own. The Assistant can draft and suggest, and it can create or update things in your family group when you ask it to. It cannot:

• buy anything, change your subscription or spend money
• delete your family's content
• invite, remove or change the permissions of any member
• send a message, email or notification to anyone outside your family group
• change a date of birth or a child's privacy settings
• connect or disconnect a calendar or mailbox

Anything with consequences needs a person to confirm it.

It refuses some things. The Assistant is configured to decline requests for content that would breach our Acceptable Use Policy, and to decline the categories of advice listed in section 4 rather than attempt them.

Teen accounts get more. See section 6.

Isolation between households. Each request carries only the asking family group's context. There is no shared context between households.

Minimisation. We send the model the least it needs to answer. Home addresses are never included.

Rate limits. Reasonable use limits apply, to protect the service and to make automated abuse impractical.

No training, no retention at the provider. Contractually committed, on zero-retention terms.

Almost nobody reads your conversations. Nobody at Northern Tech Hub reads them as a matter of routine. A named member of staff may look only where you ask us to investigate something, where we are investigating a report of abuse or a security incident, or where the law requires it. Every such access is logged.

6. Teen accounts

Where a family group is on Pro and the owner has enabled it, a teen account can use the Assistant, with these differences:

• it will not discuss topics we have judged inappropriate for a young person
• where a conversation suggests a young person needs help, it will point them towards a trusted adult and an appropriate support service rather than trying to be that support itself
• it will not produce content about other family members at a teen's request
• a teen's conversations are used no differently from anyone else's: no training, no profiling, no marketing

The account owner can turn the Assistant off for a teen account at any time, in Settings.

A note for young people. The Assistant is not a person and it is not a counsellor. It will sometimes get things wrong with complete confidence. If something is worrying you, a real person is better. Childline is free on 0800 1111, any time.

7. It will sometimes be wrong

Large language models generate text from patterns. They produce answers that are fluent and confident whether or not they are correct, and they sometimes invent facts, dates and details entirely. This is a known property of the technology, not a bug we are about to fix.

So: check anything that matters. Do not act on a date, an amount, a dose, an instruction or a deadline the Assistant gives you without checking it against the original.

We do not promise the Assistant will be accurate, available or fit for any particular purpose beyond the general standard that applies to the Pro plan under our Terms of Service and the Consumer Rights Act 2015. Section 13 of our Terms of Service explains what we are and are not responsible for.

8. Telling us when it goes wrong

If the Assistant gives you an answer that is harmful, offensive, badly wrong, or that reveals something it should not have:

• In the app: use the report option on the answer.
• By email: support@your-family.app, telling us what you asked and what it said.

Reports about a young person's safety are looked at immediately. Everything else we aim to acknowledge within [1 working day].

Reporting a bad answer genuinely helps. It does not go into training data, because nothing does, but it does tell us where the guardrails need work.

9. If this policy is broken

Depending on what has happened, we may turn the Assistant off for an account, limit what it can do, suspend the account, or close it. Where somebody's safety is involved we act first and explain afterwards. Section 10 of our main Acceptable Use Policy explains how to appeal.

10. Changes

The Assistant will change, and so will this policy. If we make a material change we will tell you by email and in the app. The date at the top always shows when it last changed.