The Globe and Mail recently published an article on the highly relevant topic of the ethics of artificial intelligence. The use of AI in recruitment and hiring practices has typically been testy in that it re-affirms biases and leaves out minority and underrepresented groups. This has been a major point of contention when it comes to the general feelings towards incorporating AI into roles that must take into account factors like gender, race, and sexual orientation.
This article from the Globe takes a unique and different perspective, and suggests ways that AI can actually help marginalized groups in the workforce. Firstly, Burton-Bloom brings up the fact that the Covid-19 pandemic has resulted in “women and marginalized communities being ousted from the work force in record numbers” due to how these groups have been disproportionately impacted by the pandemic and the stay-at-home orders. As well, many industries typically dominated by women such as food services are still facing unemployment levels abnormally higher than pre-pandemic rates. And, although the unemployment rate for racialized workers is back at the pre-pandemic levels, it is still higher as compared to non-racialized workers. In fact, “youth, Indigenous people, women with children under 6” face similar disparities. All of this raises the question of how to increase levels of employment for marginalized groups in the labour market. The article suggests that AI could be a solution.
As previously mentioned, a major concern with using AI in hiring practices is that algorithms can “reproduce the unjust biases present in our society,” as reaffirmed by Dr. Ganapini. However, Dr. Ganapini also notes that it is possible to counteract this with a “value sensitive design strategy” that is created “in accordance with our values” that allows stakeholders to embed a system with said values and then “find ways to check whether the system behaves ethically”.
Diversio Global from Toronto has actually developed an ethical AI platform that enables companies to measure and improve their diversity, equity, and inclusion practices. The platform collects personal and identifying data through an anonymous survey and combines it with other information about the company to uncover “what equitable representation looks like”. The platform then uses algorithms to identify pain points within the organization to shed light on organizational biases and inequities. With this information, the company can actively suggest solutions as a whole or within individual departments.
Although tools like that developed by Diversio Global are useful ways of utilizing AI capabilities to help address the problems of bias and underrepresentation in the workplace, it of course does not solve the issue entirely, and is not a one-size-fits-all for every type of organization.
Read the full article from the Globe and Mail.