The challenge
Balancing privacy with usability
Organisations across healthcare, finance, energy, education, and transport sectors possess valuable data critical to improving operations and services. But sharing sensitive information like health records, mobility data, or student results poses significant privacy risks.
Traditional privacy-preserving technologies often degrade data to the point of limited usability. These methods apply blanket noise to all fields, disregarding specific trust levels, sensitivities, and analytical needs of the receiving party.
Our response
Tailored privacy for trusted sharing
OptimShare, developed by CSIRO in collaboration with the Cyber Security Cooperative Research Centre (CSCRC), enables organisations to share critical and sensitive data by applying trust‑level privacy controls that preserve analytical value.
By introducing Controlled Partially Perturbed Non-Interactive Data Sharing (CPNDS), OptimShare identifies which parts of a dataset carry higher privacy risk and selectively protects them, tailoring the level of protection based on data sensitivity and the recipient’s level of trust.
OptimShare uses:
- Differential Privacy: to add statistical noise to data to prevent identification of individuals while allowing useful insights.
- Fuzzy Logic: to dynamically adjust privacy settings based on context and need.
- Cell Surprise Factor (CSF) and Personal Information Factor (PIF): to identify data points at higher risk of re-identification.
The results
Preserving privacy while unlocking potential
OptimShare has been used to protect and analyse datasets of national importance. NSW Government agencies used an early version of OptimShare (PIF) during the COVID-19 pandemic to assess and release critical case data, such as new cases of the virus, without risking individuals’ identities.
WA Health and WA Transport piloted OptimShare to explore privacy-preserving ways of sharing sensitive health and mobility data across departments and with third parties, with the goal of improving service delivery while maintaining compliance with privacy regulations.
Industry feedback has highlighted its ease of use, customisation, and ability to balance privacy and utility.
Applications of OptimShare by sector:
- Health: Share de-identified patient data between hospitals to improve care coordination and research.
- Research: Share de-identified research data across institutions without compromising confidentiality or participant privacy.
- Transport: Collaborate with city planners using anonymised GPS data to optimise traffic systems.
- Education: Evaluate student programs across departments while complying with child privacy laws.
- Finance: Share fraud pattern metadata between banks without breaching customer confidentiality.
- Energy: Enable privacy-preserving sharing of household energy usage patterns for demand forecasting.
- Agriculture: Coordinate biosecurity outbreak responses while protecting proprietary farm data
Expression of Interest
Interested in transforming how your organisation shares data? Submit an Expression of Interest to explore how OptimShare can help you unlock data’s full potential.
Testimonials:
“The Personal Information Factor (PIF) tool is a privacy tool which assesses the risks to an individual’s data within any dataset; allowing targeted protection mechanisms to be put in place. PIF has been released on GitHub as open source. In addition, the OptimShare tool is being developed and trialled by the Government of Western Australia.
Peter Bouhlas, Chief Information Security Officer, Department of Premier and Cabinet, Office of the Digital Government (DGov)
This tool, an advancement on PIF, aims to remove the manual process by automatically correcting identified risks to a defined tolerance. The Government of WA may use the tool to meet future policy requirements. OptimShare is anticipated to provide additional assurance when sharing data sets between agencies and third parties thereby protecting citizens privacy.”
Ian Oppermann, ServiceGen Co-founder, Industry Professor at UTS, Former NSW Chief Data Scientist (2015 - 2023)
“OptimShare is another Australian and world first. It was (is being) recognised as a finalist for the iAward 2023 NSW. Its areas of novelty were also sufficient to allow for an international patent to be filed. Beyond NSW government, the Western Australian (WA) Government is exploring the implementation of OptimShare for privacy-preserving data sharing in health services (WA Health) and transport (WA Transport) within the state.
OptimShare has received highly positive feedback from industry partners such as NSW Transport, NSW Education, NSW Customer Service, NSW Privacy Commission, and Revenue NSW Revenue. OptimShare offers (easy to use) a standalone version and a web-based solution for increased accessibility and flexibility for all users.
The team at Data61/ Cybersecurity CRC took on an “impossible” challenge of general protections for data sharing at scale. Whilst that challenge remains, the team have demonstrated real and meaningful progress towards elements of data sharing for row-and-column data sets. Based on the frameworks developed, more general results can be expected soon.”
Awards
OptimShare (formerly PIF tool) has received multiple accolades at both state and national levels through the prestigious iAwards and AIIA programs:
- NSW iAwards 2022: Merit Award (for PIF Tool, early version of OptimShare)
- National AIIA Awards 2022: Finalist (for PIF Tool, early version of OptimShare)
- NSW iAwards 2023: Finalist (for OptimShare)
- Victoria iAwards 2024: Merit Award (for OptimShare)
- National AIIA Awards 2024: Finalist (for OptimShare)