The second part of my interview with Dr Ian Oppermann, chief data scientist of NSW, covers issues around the individual’s right to privacy, the transparency of the government’s data sharing process & related areas.
To be honest, each of these subjects would warrant a substantial article. Hopefully, these snippets prove interesting enough for some of you to make the time to dig further.
Steve Mackay (SM): Some of the constructs you’re working on are not just ‘wicked’ they’re downright complex. Is it possible for more than a very elite group of people to get across issues like this?
Ian Oppermann (IO): What is important is that the approach we use and the frameworks we build are transparent. Anyone who wants to can get in and test the models. The average user just needs to know where to position the settings. This is similar to what is done in a lot of technology areas including cyber security. Few people know the inner workings of antivirus software, but many people feel comfortable with setting a slider to indicate how tightly they want the software to manage online interactions.
It’s also important there is widespread input to the development of frameworks. As frameworks get adopted, confidence that they are doing what they are supposed to will only come with time. We have had more than 100 years of automobile development to get to the level of vehicle safety and reliability we have today. Hopefully we will develop operational confidence in less than 100 years for data sharing, but it will still need to be developed over time.
SM: Does Australia have a unique context for data governance? What’s unique about it?
IO: Australia has some unique aspects including three layers of government, different commonwealth and state laws on privacy, rejection of a unique national identifier that many countries have, and a relatively low (but rising!) trust in government as we saw with the 2019 Edelman survey. These legal, political and cultural factors make Australia different to the rest of the world.
SM: Is there a line between ‘data governance’ and ‘data privacy’? Where would you draw this line? Or is privacy a subset of data governance?
IO: A lot of our recent work has been on developing governance frameworks which support privacy-preserving data sharing. That is a lot harder than it sounds, partly because all people-centric data has some personal information. The question is, even when de-identified, is there sufficient personal information in the data to ‘reasonably’ identify a person.
Eye colour is personal information, but not personally identifiable information by itself. Neither is hair colour, age, your street name and the school you went to. However, put all together, the combined data set could point to you as an individual.
It is tremendously difficult to work out exactly which combinations of data lead to that point of being personally identifiable. The work we have been doing with the ACS [Australian Computer Society] is trying to put a robust measure on that issue of how much personal information is present in a linked, de-identified data set and then use this measure to build risk frameworks.
Dr Ian Oppermann is the NSW Government’s chief data scientist and CEO of the NSW Data Analytics Centre. He has 27 years’ experience in the ICT sector and has led organisations with more than 300 people, delivering products and outcomes that have impacted hundreds of millions of people globally. He is a thought leader in the digital economy and is a regular speaker on big data and the impact of technology on society.
