00:02WA Police has been trialling this program since June, and as you mentioned, it is an
00:08AI facial recognition software that is used to compare faces that are scanned using normal
00:16CCTV cameras against a watch list of around 4,000 people.
00:21WA Police say that the people on the watch list include those accused of serious offences,
00:27people with a warrant out against them, and people who are missing people, as well as
00:33people who have been identified as a risk to themselves or others.
00:37This has been now happening for a couple of weeks, it is a trial and it is the first time
00:41an Australian police force has used this technology live, and WA Police say that it is something
00:48that they believe that other state police forces are watching quite closely.
00:53Obviously, this has raised privacy concerns.
00:56Yes, so facial recognition software and technology is well known for having potential racial biases,
01:05and that is because essentially this is an AI technology that is trained on thousands, tens
01:12of thousands, hundreds of thousands of faces, and through that process it learns how to recognise
01:18faces compared to other faces. Typically this data is, has like over-represented white demographic
01:30other faces, and so it sometimes faces issues with accuracy when it comes to demographics that
01:38aren't as necessarily well represented in that training data. Now, a similar, the same technology
01:45was trialled in the UK underwent official testing, and that testing showed that there was no
01:51statistical significant difference between demographics and gender in ideal conditions.
01:58But the issue that is raised is whether this technology is being used in those exact conditions as it is
02:05being trialled. For example, the conditions that were trialled are, you know, people going through a
02:11crowd straight on in perfect sunlight. Obviously, that's not exactly the way the world works, you know,
02:17different weather, different angles, and in particular, questions about how it might struggle with First
02:23Nations faces. That's because it seems very unlikely that this data has been, that this technology has been
02:31trained on a corpus of data that has First Nations faces in it. We asked WA police, we asked the
02:39provider,
02:40and neither said that they did. And so, as a result, there's reason to suspect, based on that, that it
02:46is
02:46unlikely that it would do as good a job as recognising those faces. What's the ramification of that? Well, if
02:53you're
02:53not as good at recognising those faces, you might fail to make matches, or you might make false matches. And
03:00so,
03:00in this case, what might end up happening is people would get wrongly recognised and accused of being
03:06on that watch list. Obviously, as we well know, that Indigenous Australians face already over-policing.
03:12So, there are real questions about what WA police has done to prepare for this very well-known issue.
03:19We, of course, have put it to them, and they have declined to say, but they've just said, you know,
03:23this is part of what the trial will be looking at.
Comments