AI does the lifting. People make the call.

Our view

Technology should give you more time for life. AI is good at the parts of work that are slow, repetitive or easy to put off: reading through everything, writing the first draft, keeping things in order. It is not a substitute for the judgement of a parent, a teacher, an engineer or a business owner.

So every Murphee product follows the same pattern. The AI prepares. A person reviews, changes what needs changing, and decides. We build the review step into the product rather than leaving it to good intentions.

Five things we hold ourselves to

1. People stay in control. AI proposes; a person approves. Consequential actions, such as sending a message, issuing a report or changing a child’s settings, do not happen on the AI’s say-so.
2. Show the working. Where AI makes a suggestion, we show what it was based on, so it can be checked rather than trusted.
3. Say what it is. AI output is labelled as AI. Fictional people are described as fictional. A scripted guide is called a guide.
4. Off until someone turns it on. Where a product is used by or around children, AI features start switched off and a parent or carer chooses.
5. Honest about maturity. If something is a pilot or a preview, the page says so, and we do not claim accuracy we have not measured.

Where the detail lives

This page is our approach, and it applies to everything we build. What the AI does inside a particular product, and what happens to your data there, is set out by that product itself, alongside its own privacy information and terms. We don’t repeat it here.

Several of our products are in pilot or still in development, so for now the best way to get the detail is to ask. Email hello@murphee.one and a person will answer.

On this website, the “Ask Murphee” guide is currently scripted, not AI. It answers from the product information on the page and is labelled “Product guide”. A person reads and answers every email sent to hello@murphee.one.

Your data and AI models

Your data is yours. We promote the ethical use of data, with you in control of it. How that works in practice is set out by each product; these are the principles they all follow.

• No training without your say-so. We do not use your data to train AI models unless you have explicitly given us the right to. If you do, you can take that back at any time.
• No profiling. We do not build profiles of you from what you put into a product.
• No marketing or advertising use. What you put into a product is not used to market to you, and it is not sold or passed to advertisers.
• Only what the task needs. An AI model is given the information it needs for the job in hand, and no more.

What AI gets wrong

AI systems can be confidently mistaken, can miss what a person would notice, and can reflect biases in the material they learned from. That is why our products show their sources, keep a review step, and never present a suggestion as a verified fact. If you see an AI output from one of our products that is wrong or inappropriate, tell us at hello@murphee.one. We read every report.

Questions buyers ask us

Can the AI act without approval?

Not for anything consequential. Drafting and organising happen automatically; sending, issuing, publishing and changing settings wait for a person.

Can we see why the AI suggested something?

Yes. Where a suggestion is based on evidence, the evidence is shown alongside it, so it can be checked rather than trusted.

Is any of this certified or independently evaluated?

Where a product falls under regulation, or needs independent evaluation or certification for the way it is used, we meet those requirements, and we do not offer it for that use until we do.

What applies differs from product to product, so the specifics are set out by each product, alongside its own terms and privacy information. Products that are still in pilot say so, and evaluation that is under way is described as under way. If you need the detail for a product today, email hello@murphee.one.