CAPS is proud to partner with HubHello on SAFE, a world-first early intervention program that uses machine learning and language processing technology to identify children potentially at risk of abuse or neglect.
The aim of SAFE is to make it as easy as possible to accurately verify and triage children at risk and provide referrals to the early intervention family services or child protection services that will keep them safe from harm.
SAFE was developed in response to the Royal Commission into Institutional Responses to Child Sexual Abuse, which found that all organisations in contact with children should have the systems in place to identify those at risk and respond appropriately.
By using SAFE, organisations can be confident any vulnerable children under their care are easily identified by staff, accurately verified by a child protection expert, and then promptly reported to the child protection system.
It significantly reduces the administrative burden required to comply with state and federal mandatory reporting requirements and child-safe organisation obligations.
SAFE analyses health and welfare data sets provided by childcare organisations through the Australian Government’s prescribed Child Care Management System.
This data is unique to SAFE and is based on the mandatory reporting guidelines in each Australian state and territory.
When providing the data to SAFE users must respond to simple questions designed to identify the level of risk and the type of abuse or neglect (including emotional, physical and sexual). These questions were developed in consultation with industry leaders on child protection, trauma and case management.
SAFE has six years of machine learning and a 96.2% accuracy rate in identifying at-risk children.
This data is then reviewed by experienced child protection workers that study each case and identify whether a child is at risk. The reviewer can then make a mandatory report for that child if appropriate, providing their case information to the appropriate state or territory child protection agency.
By default, cases are prioritised in order of risk to ensure the most vulnerable children receive the most urgent attention. This ensures better outcomes for the child and prevents anyone from slipping through the safety net.
The program AI becomes smarter with every case assessed. The more organisations that utilise SAFE the better our society becomes at recognising vulnerable children well before any harm can occur.
All data utilised by SAFE is encrypted and password-protected, with case information only available to authorised child protection personnel with current working with children and police checks. Those authorised to review case information in SAFE are routinely audited.
All SAFE data is hosted with Amazon Web Services, one of the most secure data platforms in the world. SAFE is also approved against Services Australia’s Secure Cloud Strategy for Integrated Third-Party Software standard and abides by the Australian Government’s 13 Australian Privacy Principles.