Bupa Australia has made a deliberate effort to diversify its data team in order to better catch algorithmic biases that could compromise machine learning models, according to Bupa Australia’s director of data technology, Craig Rowlands.
“Typically, when we’re pulling a model together, from a Bupa point of view, we’ve got a diverse team… with a number of different individuals from different perspectives,” Rowlands said during his keynote address at the Future of Financial Services Melbourne 2022 conference.
“Forgive me for saying it, but it won’t all be 50-year-old single white guys”.
Rowlands added that the data team’s models are consistently tested “from an ethics and privacy point of view” to ensure they “correlate with different populations”.
“So, it’s not just a singular lens but a multi-lens”.
More broadly, he said, the industry is also seeing an increasing number of models trained specifically to remove bias.
“On the basis of that, we’re actually going to become more tolerant as a society because, when you add the complexities of diversity in there, it’s actually going to be a better representation of that population.”