EDI and Digital Intelligence: An Action Plan
by Nicolas Sacchetti
To lead greater inclusive and responsible innovation in digital intelligence (DI), researchers Gita Ghiasi and Amanda Riddles present recommendations and action plans to help create a more inclusive environment that fosters innovation.
It has been based on a survey of people’s experience related to Equity – Diversity – Inclusivity (EDI) in IVADO’s academic ecosystem. One of the first questionnaires in the DI industry to look deeper into the EDI challenges within the context of institutional culture.
The presentation has been held on May 9, 2022, at the Policies Processes and Practices for Performance of Innovation Ecosystem (P4IE) Pre-Conference on Measuring Metrics that Matter. An event organised by 4POINT0.
« Digital intelligence is more than just AI. It also includes related fields, such as data science, business intelligence, machine learning and operations research, » explain Amanda Riddles. She is a research officer at the Université de Montréal, interested in the social and ethical impacts of DI in scientific research. Her co-presenter Gita Ghiasi advises IVADO’s EDI team on how to steer IVADO’s scientific evaluation system in digital intelligence toward equity and equality outcomes.
AI learns from train sets that are present. As a reflect of our societies, there will be a lack of representation of different groups that constitute the human kind. If we do not address this issue, AI will learn and continue in that matter by not taking into account the whole diversity of the living. In this order of ideas, there is a need to paint the idealistic society we want to live in. One equal, diverse and inclusive. This way, AI will be created with a lower risk of biases and prejudices, which can be harmful by leading to poor decisions.
Amanda Riddles explains that in order to manage these risks, people involved in the creation and governing DI need to come from diverse backgrounds. The inclusion of traditionally underrepresented groups and those most likely to be exploited or disadvantaged by it is a must to ensure EDI digital intelligence. (e.g. women, visible minorities, people with disabilities and Indigenous people.) The intersectionality factor makes it even harder on the disadvantage people.
« Digital intelligence is fast evolving and impacting all other industries, including democratic processes. This makes EDI in digital intelligence urgent. »– Amanda Riddles, information scientist
The recommendations are declining in three priorities. They go as follows. The first is to find talents in the underrepresented marginalised groups by mandating a professional outreach coordinator. The scientist adds that there is a need to « ensure that communications, websites and physical spaces are accessible to people with disabilities and inclusive of diverse identities hence the need to create accessibility and inclusion design plans. »
The second priority concern the institutions. IVADO recommends facilitating access to language courses (French and English, scientific and general) at low or no cost. As well as, « Initiatives should be created to help managers understand how to apply EDI principles in their work, influence team culture, and report regularly on their EDI efforts. » It is also recommending reviewing the allocation of resources and opportunities to ensure these processes are equitable and that certain groups are not consistently disadvantaged when compared to others.
The third priority concerns knowledge. The recommendations consist of « reviewing research and knowledge generation processes. To consider EDI at the start of the process, it is necessary to ensure its incorporation into new products/technologies developed. So, prioritise equity and inclusion by design. » Through education, we must « raise awareness of the importance of diversity and inclusion for achieving research and scientific excellence and as a matter of human rights, fairness and highlight how a lack of EDI in academic activities can hinder innovation. »
« IVADO would like to conduct a follow-up survey to compare results and find out if any progress is being made, » explain Amanda Riddles. She quotes Mary B. Anderson in her book Focusing on women : UNIFEM’s experience in mainstreaming: « Mainstreaming is a process rather than a goal that consists in bringing what can be seen as marginal into the core business and main decision-making of an organisation. »
As a next step, analysing research output is another thing that can be done. « We hope that work will be done to analyse research output and determine whether gender, sex and intersectional analysis are integrated into research activities, including the data being used, the algorithm and models used to analyse the data, and the deployment process, » says Amanda Riddles.
Analysing research output can also be done by continuously monitor EDI in research content, context and the relation between both. This would help understand how research content relate to the kinds of knowledge is being produced.
The last step Amanda Riddles presents is an action plan is to create an interactive and dynamic web application. The objective would be allowed people to « compare personal perspectives with those of other groups individuals. This can gain a better understanding of how their perceptions compared to members of different groups. »
This content has been updated on 2023-10-27 at 21 h 41 min.