Jeannie Marie Paterson
speaker
83 appearances
1 recordings
1 series
first heard Feb 2022
last heard Feb 2022
Jeannie Marie Paterson’s voice in public audio — every appearance, attributed to the second.
Trend
recordings per month · last 12 monthsNo recordings in the last 12 months.Older appearances are listed below; set an alert to hear about the next one.
Appearances
Thank you for having me.
Yeah, it's a strange new world.
What we're seeing is some agencies that provide credit rating are now trying to use algorithms to help with that credit score.
And that means they provide all sorts of data to the algorithm, which then comes up with the score.
And one of those factors that is being considered is postcodes.
Now, the thing you mentioned is you might be told your postcode's not good enough.
The point here is people who are trying to borrow money won't even be told their postcode was considered.
They'll just be told they're not a good borrower.
That's part of the problem, that the algorithms that are being used are grabbing all sorts of data that we don't really know about, which may not be that relevant, which may not be fair, but is still potentially affecting whether a person gets a loan.
No, where you live does not affect an individual's capacity to repay a mortgage.
But what these algorithmic processes are doing is looking at groups of data, groups of people that look like each other or live near each other or have similar profiles.
So that if the algorithm finds that in a particular postcode, either people are bad at repaying their loans or equally people aren't borrowing money at all, then the algorithm is going to say, well, that postcode correlates with poor ability to repay.
And that tells you nothing about a particular individual's ability to repay.
And it might not be relevant at all.
It might just be people in that postcode don't borrow very much money.
It can be a form of discrimination because it can be a correlation.
The postcode can also be a correlation with people who are older and
or younger or of a particular cultural background or ethnic background and then we have what's called discrimination by proxy the score is not based on somebody's race but it's de facto based on race because it's picking up where people of a particular cultural or ethnic background live and in fact there's been a lot of problems in the us with this practice
That's right.
So automation is justified as being more efficient.
Showing 1–20 of 83 · page 1 of 5
Next →