How We Use AI

We are open about exactly where AI appears in SpotCheck, and more importantly, where it never does. The short version: AI writes the words. It never writes the numbers.

1 Written summaries in your report

Some sections of a report open with a short plain-English summary. These are drafted by an AI language model, but only after your report's data has been fully computed. The AI is handed the finished numbers and asked to describe them in sentences. It is never asked to produce a figure, estimate a value, or fill a gap. Automated fact-checking routinely compares these summaries against the report's underlying data.

2 Report translations

If you read a report in one of our 47 supported languages, the translation is performed by an AI language model. It translates the wording only. Every number, chart, and map comes straight from the original English report.

3 The data is never AI-generated

Every measurement, rating, statistic, and map in a SpotCheck report comes from our data pipeline: government environmental agencies, national statistics offices, police records, transport authorities, and our published methodology, applied identically to every address. Generative AI never creates, edits, estimates, or fills in a data point. If a value is not in our sources, the report says so. The AI is never asked to guess.

4 What about the modelling?

Some metrics, such as air-quality estimates between monitoring stations, use machine-learning models trained on measured government data. These are statistical models, the same kind used in weather forecasting. They are not chatbots and not generative AI. They cannot invent facts, and they run identically for every property, every time.

AI writes the words, never the numbers

Every fact in a SpotCheck report comes from verifiable data sources and one consistent methodology. AI helps us explain that data clearly. It is never allowed to create it.

Questions about how we use AI? Contact us at [email protected]