By Monica Toro Ospina and Jawad Lasri, Anti-Racism Committee members

The faster that artificial intelligence (AI) evolves, the clearer it becomes that its systems are far from neutral: it can reproduce, even accentuate social biases, leading to racial discrimination. These biases reflect and amplify the prejudice in society, as can be seen in several examples of sexist and racist discrimination. It is fed in particular by stereotypical data and a lack of diversity in development teams.

Black and Afro-descendant communities remain largely under-represented in this sector, which contributes to reinforcing existing discrimination and can have real consequences, such as the exclusion of women or racialized people. The gravity of the situation calls for urgent action to make AI more ethical, inclusive and equitable, focusing on diversity, collective vigilance and technology that truly serves everyone.

On that note, a report published in 2016 revealed that AI software in the US judicial system reinforced racist biases by overestimating the risk of repeat offences among Black defendants. This case clearly illustrates how AI can exacerbate discrimination when it is based on biased data. However, if used responsibly, AI can also become a powerful tool for detecting and correcting injustice. For example, tools such as Ravel Law can analyze court decisions and highlight potential biases. In Canada, AI could also address the lack of racial data and further promote transparency and fairness in the legal system.

In light of these challenges, it is essential that we promote transparency, encourage diversity in the AI sector and implement strict regulation to guarantee fairness and respect of fundamental rights.

To this end, a project was developed in partnership with the Observatory of Black Communities in Quebec. This project aims to create a collaborative space to promote the participation of these communities, strengthen their impact and stir innovation in the field of artificial intelligence.