Mitigating Artificial Intelligence Bias in Mental Healthcare: Promoting Equity for Racialized Canadians

Technological integration provides an effective solution for ensuring racial equity in healthcare. With growing optimism for AI integration in mental healthcare, an intersectional approach is required to ensure that patients of all racial groups benefit from the advantages AI algorithms are anticipated to provide. Despite the long-standing knowledge that racialized groups in Canada encounter challenges in attaining equitable mental healthcare, little has been done to mitigate these inequities. Systemic racism and Canada’s 'colour-blind' healthcare policies have perpetuated inequities in mental healthcare access. Given the anticipation around this emerging technology, it must be acknowledged prior to its use in the mental healthcare space in Canada that data which is representative of racialized groups is limited due to sparse research. Certain studies propose that if this unrepresentative data is used to train AI models, inferences made may exclude racialized groups due to inherent AI bias, further exacerbating existing mental healthcare disparities. 

This study employs qualitative and quantitative methodologies that provide a link between equitable access to mental healthcare for racialized groups in Canada and AI’s ability to provide equitable mental healthcare access to racialized groups despite its inherent bias. The qualitative aspect utilizes a narrative review of existing literature. The quantitative method includes augmenting data to present how racially inclusive and representative data can be created, possibly reducing inherent AI bias. The findings demonstrate that inherent bias may occur at any stage of algorithmic development, and mitigating bias demands an approach beyond addressing the historical absence of racially inclusive mental health data. The findings further explore strategies to mitigate inherent AI bias. Notably, inclusive in these strategies is the reweighing of existing data, which will be demonstrated to achieve data representative of the mental health prevalence of racialized groups in Canada. 

As this is the first research study to focus on linking the topics previously discussed, it contributes a unique viewpoint on achieving a more equitable future for all mental healthcare patients in Canada.