Earlier this year, a study led by MIT Media Lab uncovered gender and racial bias in the algorithm behind Amazon’s facial recognition software. Amazon’s Rekognition, marketed to law enforcement – among others – in the United States, proved less than accurate when it came to identifying someone’s gender if they were darker-skinned, with tests showing that it misclassified darker-skinned women 31 per cent of the time; by contrast there was a zero per cent error rate for lighter-skinned males. (Amazon disputed the study’s findings.)
As AI-powered technologies such as facial recognition go mainstream, debate is raging as to how they should be deployed in the real world. From smart voice assistants and medical diagnostic tools to services such as Netflix offering personalised recommendations, AI is already stitched into our daily lives. But whether we’re talking about its malign influence on elections, the flaws in crime-fighting by algorithm, or how it is open to abuse by authoritarian governments, it’s clear that by their very nature great technological leaps forward often carry with them unintended consequences – ones which can quickly flare up into full-blown crises that hit organisations in the bottom line, as well as society at large.
This report zeroes in on four such challenges – where businesses are buffeted by new technology – and considers how the C-suite can lead their companies in response.