News from the open internet

Culture

Women leaders at PepsiCo., the NFL and Spotify speak out on breaking down AI biases

Women stands on a series of blue building blocks and looks into the distance.

Illustration by Holly Warfield / Getty / The Current

Just like humans, artificial intelligence is prone to bias.

Ivana Bartoletti, the global chief privacy and AI governance officer at consultancy Wipro and co-founder of the Women Leading in AI Network, recalled one example at last week’s inaugural Women in AI Summit.

When ChatGPT was asked to create a conversation between a boy and a girl about what they want to study in school, Bartoletti said, the boy stated: “I’m going to study engineering,” while the girl said “I’m going to study art because I do not understand numbers.”

The example served as a call to action for the women gathered at Webster Hall in New York at the summit presented by nonprofit The Female Quotient and digital media company ATTN. Speakers from Chelsea Clinton and Ashley Graham to leading AI experts like Bartoletti discussed the ethical ramification and practical business applications of AI. One alarm bell they rang: If women aren’t in the room for the creation and regulation of AI programs, the technology will be flawed and perpetuate bias.

Already, there’s a gender gap in AI roles, with women representing only 26% of the workforce globally, according to a 2020 World Economic Forum report.

Brands like Dove, Thinx and Sephora have created entire campaigns about AI’s tendencies to stereotype women. Similarly, a video for the event underscored the problem, with an AI-generated image of a female astronaut, complete with supermodel looks and eyelash extensions.

“Our rights, our values, our power, our interests are simply going to be written out,” warned Reshma Saujani, CEO and founder of nonprofit Moms First and founder of Girls Who Code.

“As we speak today, seven days after the election, there is nothing that we should be thinking about other than the fact of how we should maintain our power, our relevance, our voice. And that means if there’s a tool — AI — that is rising in prominence, that is going to disrupt everything, including the rest of our world — you better believe we better be sitting around the table,” continued Saujani. “We have to take it like we have to take everything.”

Leading from here

Even with the constraints in place, women in leading roles at major brands like Pepsi, the NFL and Spotify are working to make AI more inclusive.

On one of the panels, Umi Patel, VP of consumer insights and analytics at PepsiCo Inc. North America, said the brand is working with AI and “empathy partners” like Discuss.io and Voxpopme to surface deeper consumer insights around perceptions of products like iced coffee. Not only does PepsiCo collect data but it then films certain surveys to see how some customers arrive at their takeaways, adding a human element to AI data collection and analysis.

“The data will tell you what people are doing, not why,” Patel said on the panel about AI and storytelling.


Meanwhile, the NFL is using AI to create a platform where people can find where their favorite games are playing on a daily basis.

“I can speak for probably everybody that when you go to find an NFL game, it’s not necessarily cut and dry — it could be on Fox or CBS or ESPN. So we used audience data to build out a ‘ways to watch’ platform,” said Ashley Downey, senior director of marketing at the NFL.

And this, The Female Quotient founder and CEO Shelley Zallis pointed out, can help people discover more women’s sports games, further driving up its popularity after the Taylor Swift effect.

For Spotify, which has used AI to help with discovery on the platform through an AI DJ called DJ X, continues to have trouble with it recommending men over female artists, said Brittney Le Roy, global head of product and technology communications at Spotify on a panel.

“How do we ensure diversity in the mix? It goes back to the algorithm,” said Le Roy. “You have to train the machine.”

What’s clear is that every AI machine is only as productive as the women steering the way.