How to Foster Inclusion & Trust in Artificial Intelligence Work Groups

Women in AI (WAI)
4 min readJul 30, 2019

by Elizabeth M. Adams

https://www.linkedin.com/in/lizadams/

Interested in accelerating innovation, increasing value, delivering Trustworthy Artificial Intelligence (AI) while ensuring diverse work groups participate in a meaningful way? Fostering an environment of inclusion and trust is your answer.

In September of 2018, the following article was written Building a Diverse and Inclusive AI Workforce. It states “The value of a diverse and inclusive AI workforce cannot be overstated. AI teams with diverse backgrounds, perspectives, and talents contribute unique skills, ask new and different questions, and cover each other’s blind spots. And yet, research shows that the AI field doesn’t adequately reflect the broader population, which suggests that globally, we’re missing out on the value that diverse teams bring to AI development, implementation, and research.”

Today we are still experiencing a lack of diversity and inclusion (D&I) in AI work groups. As a D&I practitioner and technologist, I’m a strong advocate for using creative and agile solutions in the short-term as building blocks for future solutions.

Organizations win by bringing in diverse talent through unconventional means until a more formal process to attract and retain a diverse AI workforce is in place. Unconventional ways to find diverse talent include connecting with social media groups, consultants, contractors, educators, bloggers, non profits, think tanks and community leaders to name a few. Many are willing to assist and a quick search on LinkedIn gives you an idea of the depth of talent available.

How organizations can foster inclusion and trust in AI work groups.

Adopting an environment where diverse perspectives are valued is key.

Here are a few ways to get started.

1. Allow each group member (staff, consultants, vendors, customers) time to share why & how each person is connected to the project and what it means to them and their work. While this activity may be time consuming in the beginning, it creates interest and respect for the time and expertise each person brings to the table.

2. Lead with an emphasis on the importance of adopting a learning mindset. Give people real-time permission to offer feedback respectfully. Modeling how people should give and receive feedback, will help work groups to not take feedback personally. This also creates a positive path for members to work together on future efforts.

3. Consider allowing non-technical, non-experts and/or invited diverse work groups to talk first when discussing requirements, ideas and or concerns. When those closest to the work talk first, discussions tend to be less inclusive as others don’t feel comfortable speaking up.

Additional Opportunities to Foster Inclusion & Trust

4. Set aside time during group work to pause and ask what might be missing. This gives people permission to voice questions or concerns without feeling like their opinions are invalid.

5. Give time for groups to create task and goal interdependence among members emphasizing the necessity of all roles. People will take more risks and express opinions when their role has been communicated as necessary for the success of the effort.

6. Focus time on team building so that individuals can discover similarities and create interpersonal connections. This breaks down personal barriers and allows for a more meaningful and respectful professional work experience.

7. Consider designating someone as the official AI inclusion advocate to ensure members feel comfortable speaking up when something has been left unsaid. This also helps people recognize when their communication style may hinder others from speaking freely.

_____________

Wrapping Up

As per the article mentioned above: “When diverse voices are left out of AI, the reliability and fairness of AI systems come into question. Evidence shows that existing societal biases including sexism, racism and other forms of discrimination are being built into the machine learning algorithms underlying AI technology.”

I’ve met with experts from all over the world to better understand their perspectives on diversity and inclusion in AI. I’ve spoken with researchers, CTO’s, data scientists, attorneys, STEM experts, product managers, marketing teams, nonprofits, front line managers and C-Suite executives. All agree diversity and inclusion is needed and is the heartbeat of Trustworthy AI.

Yet… they don’t know how to create an environment that fosters inclusion and trust. They don’t know how to speed innovation, accelerate value and ensure diverse work groups participate in a meaningful way. Without tools to build inclusive AI work groups, D&I will be an afterthought. The good news is that solutions are readily available. Technology leaders should work more closely with D&I leaders.

AI working environments should include diverse teams, processes and procedures where inclusion and trust are non-negotiable standards.

_____________________

Elizabeth Adams works in technology integration at the intersection of Cyber Security, Data Ethics and AI Governance. To ensure AI innovation continues as companies, countries and governments move to Trustworthy AI, Elizabeth advises on AI frameworks with a focus on positive social impacts.

Elizabeth is a Diversity & Inclusion in Artificial Intelligence practitioner, advisor, consultant, speaker, writer and author. She has facilitated Diversity & Inclusion learning events focused on racial and gender bias in Facial Recognition Technology, Video Surveillance, Predictive Analytics and Children’s Rights.

Elizabeth combines her professional experience of 25 years in STEM, with her community experience on the Racial Equity Community Advisory Committee for the City of Minneapolis, and her D&I Certificate program from Cornell University to bring a unique perspective to the AI for Social Good conversation.

--

--

Women in AI (WAI)

The first Global community of Women experts and influencers in AI.