Data science technologies, pioneered in the private sector, are now ripe for transforming the public sector. However, both government policy and technology providers need to address two pressing public concerns: DIGITAL TRUST (privacy and security) and PERSONAL DATA (ownership and beneficial exploitation).

The impact from ‘smartification’ of public infrastructure and services will be far more significant in comparison to any other sector given the government’s function and importance to every individual and institution. Potential applications range from public engagement through natural text and speech Chatbots, to providing decision support for civil servants via AI-based Robo-advisors, to real-time management of the public infrastructure through the Internet of Things and blockchain, to securing public records using distributed ledgers, and, encoding and codifying laws using smart contracts.

However, in many cases current uses of automated decision-making systems have been shown to cause adverse impacts on important life events of individuals – examples range from bias in recruitment of job-applicants, to credit scoring in loans and insurance, and to sentencing of criminals.

Also, state surveillance and manipulation of voter behaviour have become the early examples of how such developments may amplify the asymmetry of power (between citizen and those utilising such technologies) causing severe damage to the democratic processes. The Bitcoin ‘hype’, with its correlating energy usage, has also shown the environmental cost of the highly complex computations, as well as indicating other potential unpredicted and unintended consequences.

On the other hand, the cost of not using – or the slow uptake of – data science technologies in the public sector is also potentially huge, given that all other aspects of our lives are changing fast under the ongoing digital revolution. It then follows that the stakes could be much higher in both the use and the avoidance of these technologies for public decision making and service delivery. This will require a careful cost/benefit analysis before implementation at scale.

The fourth conference in the Data for Policy series therefore highlights ‘Digital Trust and Personal Data’ as its main theme. The conference will also welcome contributions in the broader data science for government and policy discussions. In particular, submissions around the value and harm of using data in the public sector, deployment experience in government, ‘digital ethics’ and ‘ethics engineering’ concepts, personal data sharing frameworks and technologies, transparency in machine learning processes, analytics at source, and secure data transaction methodologies are encouraged.