Challenging robo decisions a safeguard humans may need: expert

Colin Gavaghan says there is growing concern about the lack of transparency about decisions made by computer algorithms. Having a human in the loop won't be enough to ensure decisions made using artificial intelligence will always be fair, says an Otago University academic who is helping shaping an emerging international debate on digital rights. Computers have already been used in New Zealand to assist in parole decisions by forecasting the chances of individual prisoners reoffending, and to identify ACC claimants who might be staying on benefits longer than expected. Software has also used by the Social Development Ministry in a brief and controversial experiment to predict children at risk of abuse. But Professor Colin Gavaghan warns an explosion of interest in artificial intelligence (AI) could lead to half-baked software flooding the market which won't be good at what it is supposed to do.