As budgets become leaner, councils are increasingly looking to new lower-cost ways to protect children from abuse and intervene early to prevent serious incidents. They have turned towards developing predictive analytics that use algorithms to determine families that may need the attention of child services. Yet, this is not a definitive solution because it is potentially a breach of privacy and data rights, especially with this year’s introduction of GDPR.
The Guardian has uncovered that at least five local councils, including Bristol, Hackney and Newham, have incorporated 377,000 people’s data into predictive systems with the aim of identifying child abusers. So far, the algorithms have generated 350 risk alerts for families in Hackney, while the Thurrock council programme produced 300 alerts, indicating early success for the project.
Hackney and Thurrock councils have hired he private firm Xanture to develop these systems, giving an external company access to school attendance records, police records on anti-social behaviour and domestic violence records.
As with all systems that use people’s personal data there are serious concerns about the protection of sensitive information, especially where children are concerned. Under GDPR, organisations must provide an option of consent in an intelligible and easily accessible form. Councils must adhere to these laws, and in this context, it is unclear whether GDPR has been fully adhered to. These councils have also faced criticisms because an automated system may perpetuate minority discrimination and stereotyping.
The Information Commissioner’s Office responded to the the Telegraph by saying: “We would expect any council using such technology to have fully considered the privacy risks, including conducting a thorough Data Protection Impact Assessment, and to have taken steps to address those risks.”
Despite privacy and data issues, advocates claim that predictive analytics is an effective way to use limited council resources to counter child abuse and act before tragedies happen. By 2020, local governments will have to face £16bn worth of funding reductions from central government, so councils need to act now to develop new approaches to providing services with less manpower and reduced costs. Although not perfect, this scheme is an important step towards increasing reliance on analytics to determine the allocation of government services, which reduces demands for staff and in turn costs.