Algorithms, diversity and the tech industry – why we need to get this right!
I recently watched an excellent ted talk by Joy Buolamwini regarding her fight against algorithmic bias and why it matters. Put simply, her talk considered the issues arising from a lack of diversity in tech e.g. that algorithms can exhibit unintended racial or gender biases of the largely White, Asian and male population of the tech industry.
Since we are in the midst of a fourth industrial revolution – with artificial intelligence at the core – we need to be able to trust in the veracity of the data produced by algorithms. To minimise the risk of unconscious biases infecting the coding of algorithms, we need to ensure that there is diversity in the tech space.
James Damore (formerly of Google) complained in his now infamous memo that the focus on diversity in the technology sector is creating an “ideological echo chamber” and that Google’s programmes to increase diversity are politically driven and divisive. For instance, he referred to the fact there are biological differences between the genders which can provide a non-discriminatory explanation for the lack of female representation in tech.
In my view, it is imperative that tech companies focus on diversity so as to ensure that algorithms are fit for purpose. I do not want algorithms created to help sift through prospective job applicants to reflect the biased views of those who developed them e.g. weeding out women or those from certain ethnic or class backgrounds because of their perceived inferiority and I am sure I am not the only one. So while I am optimistic about the role AI will play in the future, we cannot ignore that algorithms will simply reflect the views of those who create them. Therefore, it is of utmost importance that those who code reflect all of us – not only on the grounds of race and gender but on any other characteristics which are likely to shape how we interact with the world such as disability, age, class, sexual orientation, religion or philosophical belief.