The creation of a new group to advise on AI and technology product strategy by Facebook parent Meta Platforms has been criticized by an executive member of the Responsible AI Institute (RAI) for what on the surface appears to be an acute lack of diversity.
“It was definitely striking, I will say,” said Alyssa Lefaivre Škopac, head of global partnerships and growth with the RAI, of a panel that is made up of four white male entrepreneurs.
“We are hearing a lot of feedback from our community that this is problematic,” she said. “The diversity question is important to be able to make sure that there are no blind spots or areas where bias is not easily identified when you are all coming from the same perspective.”
The Meta Advisory Group, Meta states on its website, is “composed of outside advisors that Meta’s management team will periodically consult with on strategic opportunities related to our technology and product roadmap.”
The four are Patrick Collison, co-founder and CEO of Stripe, a company that builds financial infrastructure for the internet; Nat Friedman, an entrepreneur and investor who specializes in infrastructure, AI, and developer companies; Tobi Lütke, the founder and CEO of Shopify; and technology investor Charlie Songhurst.
Škopac said the formation of the advisory group does not look to be a good business practice: “I don’t think that they’re going to have the right diversity of perspective to give them the input on their product roadmap and strategy that they’re looking to seek when not only are people coming from the same kind of technological perspective, but they are all white men.” She also pointed to the similarity in the group members’ backgrounds: “I am not seeing representation from civil society, or other types of ecosystem partners as opposed to just technology companies.”
It’s a wake-up call for any organization on how not to set up such an advisory group, she said. “They need to be representative of a variety of functions so it cannot just be all from one business unit. We need to see compliance, legal, HR, procurement and technology teams and they have to be diverse in the type of function they do and definitely diverse in perspectives. We need to see a balance of perspectives from race, gender and all those types of things, and it also has to be a thoughtful exercise.”
It is paramount, she said, that members of councils or internal groups created to govern AI actually have stature to make decisions and authority to direct things as they see fit from a responsible AI perspective.
Advancing AI
A Gartner research paper released earlier this month, How to Advance AI Without Sacrificing Diversity, Equity and Inclusion, noted that the rapid uptake of AI, along with biases inherent in AI models used, is affecting enterprise diversity, equity, and inclusion (DEI) initiatives, skills development, and talent acquisition, and ultimately undermining workplace diversity.
“Diversity in AI teams is crucial due to the pervasive issue of biases within AI systems. Unintentional biases, particularly gender, confirmation, and racial bias, can impede the effectiveness of AI deployment and pose regulatory and reputational risks for organizations.”
Gartner analysts David Sugden and Austynn Eubank, authors of the paper, called on CIOs to reassess existing DEI strategies to prevent generative AI from exacerbating biases.
“Utilize tools and frameworks like AI trust, risk and security management (AI TRiSM) to unlock the potential of AI while reducing risks of potential bias to enhance the likelihood of inclusivity and equity. Address the digital dexterity gap by implementing thoughtful training to foster a symbiotic relationship with AI and equip the workforce with necessary skis for technological advancements,” they wrote.
Škopac said that “if you do not have people with diverse perspectives in the process of building and developing and using these systems, they are going to have a severe risk of perpetuating bias.
“Nothing is a foregone conclusion, but it has been proven time and time again that, without diversity in the development of AI systems, bias can be systemically perpetuated. When we see these large tech companies putting together advisory boards, it is something to certainly keep an eye on and think about. It was definitely disappointing to see the advisory panel from Meta take the composition it did.”
