Computerworld

New Zealand government’s use of algorithms under scrutiny

Project assessing AI headed by CEOs of Stats and Internal Affairs
  • Sarah Putt (Computerworld New Zealand)
  • 24 May, 2018 14:39

A project examining the use of algorithms by NZ government agencies has been announced by the government digital services minister Clare Curran and statistics minister James Shaw.

It will be led by the chief executive of Stats NZ Liz MacPherson and the chief executive of the Department of Internal Affairs Colin MacDonald, with the first stage of the review expected to be completed by August. 

“The government is acutely aware of the need to ensure transparency and accountability as interest grows regarding the challenges and opportunities associated with emerging technology such as artificial intelligence (AI)”, says Curran.

“Jurisdictions around the world are looking at how their data and privacy laws are fit for the digital age, with examples such as the Privacy Bill in New Zealand and the General Data Protection Regulation (GDPR) in Europe, which comes in to effect on Friday.”

Curran says the project is part of a programme of work which includes the recently announced AI action plan, and international work with the Digital 7 nations - UK, Israel, Estonia, South Korea, Canada and Uruguay – on digital rights.

Shaw says that examples of algorithms currently in use include computer programs used by the Ministry of Health to ensure donated organs save lives, or the NZ Transport Agency’s computer modelling to make roads safer.

“They show the power of data to make a positive difference to New Zealanders. But there are challenges as well, and we need to ensure that transparency and procedural fairness are maintained,” he says.

Debate on the rise of algorithms and the impact they have on the decision making that effects New Zealand citizens was the focus of a recent Privacy Forum in Wellington.

Associate Professor Colin Gavaghan from Otago University, who is researching how AI is handled in a number of jurisdictions, says an area of debate is if it’s possible to design models in which the decisions that are made by algorithms can be explained. That is, how can you maintain transparency when the algorithm is straightforward at the start but once it has been applied to data, it ‘learns’ and changes, and as a result its decision-making process may become opaque?

Ministry of Social Development deputy chief executive Nick Blakeley said the Ministry is looking at how it can be more open about how it uses people’s data. “It’s not just putting an algorithm on the website, you have to explain how it is used, what is the business setting, is a human in the loop?” he says.