Our principles
How we use AI
Information about a child's development is among the most sensitive data there is. We hold that fact alongside another one: families are waiting years for understanding and help, and AI, used carefully, can change what's possible for them.
These are the principles every AI feature at Assembly is built and reviewed against.
Humans in the loop
Clinicians design our methodology and review what we publish. AI helps us tailor and explain, it doesn't make decisions about your child on its own, and it doesn't diagnose.
Privacy by design
We start every feature from the question "what's the least data this needs?", not "what could we collect?". Sensitive details stay out of places they don't need to be, including our own analytics and logs.
Your data stays yours
Your family's data belongs to you. We don't sell it, and you can ask us what we hold and have it deleted. The platform runs in the UK: hosting, storage and AI processing all happen in Microsoft Azure's UK South region.
Never used to train AI models
Your family's data is not shared with third parties to train their AI models, and we don't use your child's data to train our own.
Grounded and inspectable
When our AI gives advice, it's anchored to clinical guidance and your child's profile, not to whatever the internet says. The report shows its workings in plain language.
Benefit that earns the risk
We weigh every use of AI against one question: does it materially help a child sooner? If a feature doesn't clear that bar, it doesn't ship.
For healthcare services and schools
If you're evaluating Assembly for the families you support, we maintain a Data Protection Impact Assessment for our processing and are working through the NHS Digital Technology Assessment Criteria (DTAC). We're happy to walk your information governance team through our security and AI position, info@helloassembly.com.
For how we handle personal data generally, see our privacy policy.