AI IRL Podcast Episode 40: The Cultural Intelligence of AI with Davar Ardalan
True or false? Culture and AI aren’t that connected.
At first blush, you may think that AI doesn’t have many cultural aspects, that it’s all technology.
Whereas AI is directly informed by our familial, regional, national, and historical cultures.
“I absolutely believe that our interactions with chatbots will become second nature,” Davar said.
“There’s room for other people. It really can’t just be technologists who are solving the future of AI.” —Davar Ardalan
AI Is an Opportunity
Davar recently soft-launched a data set challenge, an algorithm for historic women. “The opportunity is that we look at AI systems and understand that they for the most part do still carry the biases that were created when their developers created them,” Davar explained.
But not to complain about — to take control of and shape ourselves.
Collecting stories of women in science, technology, history, and folk culture is just one step.
Bringing bias out from behind the curtain of design is another major milestone. It’ll take acknowledging the problem and bringing light to solutions.
An opportunity that makes Davar excited is how brands are messaging through chatbots. “They’re still falling short in the ability to relate to people in a more human, personal way,” she pointed out.
Chatbots are just screaming for the ability for conversational interfaces to engage in a more culturally relevant way.
“Cultural intelligence is the missing key because we can’t only use the standard metrics of demographics, environmental, economic data to build AI systems.” —Davar Ardalan
AIs Need Cultural Intelligence
Connecting with consumers in a more authentic and meaningful way are going to rely on AI.
“Cultural intelligence is the missing key because we can’t only use the standard metrics of demographics, environmental, economic data to build AI systems,” Davar said. “Cultural intelligence is going to have to become a prime metric so that people feel themselves in these systems.”
Davar’s two decades as a journalist at NPR gave her insight into how this could be accomplished.
You have to write a story that intimately reaches one person but can be amplified to 30 million users.
“It’s equally important for us to make sure that every voice is represented, even though we’re creating AI products and solutions at scale,” she said.
People’s backgrounds, values, and understanding should inform the cultural storytelling that AI really needs.
“Chatbots are still falling short in the ability to relate to people in a more human, personal way.” —Davar Ardalan
How to Increase AI Cultural Intelligence
1. Don’t believe that the AI field is just for technicals.
It’ll take diverse and interdisciplinary designers — like historians and psychologists and geneticists and journalists.
2. Give AI empathy and inclusivity.
With increased conversational interfaces, we need to commit to teaching AI how to personalize information the way humans do, from a cultural framework. They’ll need to access the next layer of detail, but 20 layers deep.
3. Acknowledge that bias is pervasive.
By the time the 2020 Census comes around, more than half of American children are expected to be part of a minority race or ethnic group.
Yet cultural intelligence in AI is low and our data will probably show bias for a while. So we should start straightening that out today.
“By the time the tenth year comes and it’s been a global collaboration to come together to get these data sets, at least we’ve built a foundation,” Davar said.
4. Understand your value.
A small, rich data set is incredibly valuable compared with a broad, sloppy one.
A company that isn’t highly technical will attract investors who value collaboration and diversity.
“There’s room for other people. It really can’t just be technologists who are solving the future of AI,” Davar said.
If you don’t use iTunes, you can also find every episode here.