Adjournment Debate - Artificial Intelligence

There are a lot of questions being asked about artificial intelligence at the moment. It's human nature to be fearful of the unknown and of change. Obviously, we need to look carefully at the implications of this technology, but, as we think about how we respond as a nation to artificial intelligence, it's useful to bear in mind some of the ways in which artificial intelligence is being used today.

The tools being used by services like YouTube, Netflix and Spotify to serve up videos or songs that we might like to see or listen to next are one good example of artificial intelligence in action, and many of us find it pretty convenient. These tools become better over time at working out our preferences. If you regularly choose eighties artists like the Thompson Twins, Alison Moyet and Frankie Goes to Hollywood, you tend to get more of these tunes served up to you—I know I do! The tools use data from previous choices you might have made. They also use data from choices made by other people. You might have chosen 'All Cried Out' by Alison Moyet. If a lot of other people who chose this song also chose 'Don't You Want Me' by the Human League, then it's quite likely that 'Don't You Want Me' will appear in your feed. The serious point is that that is but one of many examples of the way artificial intelligence is part of all of our lives, and some of the benefits it is delivering are profound.

I recently visited professor of neurology Dr Michael Barrett, a world-leading expert in the treatment of multiple sclerosis. He showed me computer images of brain scans and explained how his team had developed an artificial intelligence tool to look at brain scans and identify signs of this terrible disease. Such tools do not replace the work of radiologists, but they do allow radiologists to deal with higher volumes of scans and allow a given patient's scans to be seen more quickly.

A similar approach is being used in scanning for many other conditions in medicine. Australian start-up business harrison.ai has developed AI tools used by radiologists, and it's estimated that about a third of radiology practices are now using some form of artificial intelligence to assist in their work. This is but one of many situations where a large number of images or a large quantity of video need to be examined, and artificial intelligence tools are increasingly being used for this work.

Endeavour Energy has worked with Optus, Amazon Web Services and Unleash live on a project to use drones to fly above powerlines to identify places where the line is at risk of failing. Typically the line gets hotter when it is nearing failure, and this is captured by infrared video. The drone captures this video and sends it back to be examined by an artificial intelligence tool, which then flags locations for human investigation. This kind of preventative scanning work was simply not being done previously but is now feasible thanks to artificial intelligence.

Australian start-up company Trellis Data has developed artificial intelligence tools to scan shipping containers coming into Australia for biosecurity risks. For example, insects may have built nests on the outside of the container while it was sitting on the ground in another country waiting to be shipped. The tool examines images of the container from cameras located in the ports where the containers are unloaded from ships, and it flags containers where there's a risk. In a two-week trial period, with the cameras scanning containers from multiple angles in just four seconds, more than 1.7 million images of containers were analysed, with four per cent of containers having pests that needed to be removed. While the Department of Agriculture, Fisheries and Forestry is only able to inspect about three per cent of all goods manually, the benefit of using this artificial intelligence tool is that containers were scanned more quickly and efficiently. Rather than causing jobs to be lost, this tool allows a higher percentage of incoming containers to be scanned, and it allows staff to then have their efforts focused on a more detailed examination of the containers which the artificial intelligence tool has flagged as presenting a risk.

Similarly, in the example I gave previously, with Endeavour Energy using artificial intelligence, that system allows the skilled maintenance staff to be directed to locations where they can do preventative maintenance work rather than having to simply engage in physical inspection. As work they were previously doing is now able to be done by the artificial intelligence tool, those skilled staff are able to be deployed to other high-value and more productive work.

It's clear that many businesses are developing ways to use artificial intelligence in a growing range of settings, generally to allow existing processes to be carried out more quickly and efficiently. Rather than leading to a reduction in jobs, in many of these instances artificial intelligence is instead resulting in an increase in activity. We need to think carefully about the risks that AI might bring, but we also need to be alive to the very significant benefits we are seeing from it every day.