University and councils develop AI tool that predicts safety issues in social housing

The PRISM tool also incorporates traditional building records, such as repair histories and Energy Performance Certificates, and data the landlord has on tenants, including fuel poverty indicators and rent arrears.

When the model is in force, it will send alerts to housing officers who will make ultimately make the decisions on people’s homes or welfare. 

Dr Ramit Debnath, executive director of the Centre for Human-Inspired AI at the University of Cambridge, said the unique aspect of the tool is that it will predict safety issues not just from observation data, but also data from lived experience.

However, this is also where the project will encounter the biggest challenge. 

“Modelling a building with a machine learning model is relatively straightforward,” Dr Debnath said. “Removing all the risks around personal data and making it tight within the context of its ethics – that’s the complex part.”

Researchers said the model uses anonymised data and is designed so that nothing in its outputs can be traced back to a named individual.

They said this means that landlords do not have to wait for tenants to report issues and can intervene before things break, potentially avoiding wider damage to homes that can happen as a result.

Peter Campbell, head of housing at South Cambridgeshire Council, stressed that the tool will help find and support tenants most at risk from a safety issue, amid concerns over people who currently have little contact with the local authority.

“What we’re doing now is identifying people with whom we’ve had absolutely no contact and prioritising them for a home visit,” he said. “But we don’t have the resources to do that for everybody, all the time.”

Mr Campbell also said that making use of information to guide decisions is an increasing requirement of housing providers by authorities.

He said: “There’s been a changing approach to the way social housing is managed through the housing regulator. There’s an expectation from government to make better use of data in order to plan our services.”

The people involved in the project hope the PRISM tool can be scaled up and rolled out across councils in the UK.

News of the project comes amid a more stringent regulatory inspection regime and ahead of second phase of Awaab’s Law coming into effect this October. The strict timeframes will be applied to several more hazards and mean landlords must either fix their homes quickly or move residents out for their safety.

Alongside this, housing providers are increasingly using AI, such as for allocating temporary accommodation and boosting tenant engagement.