Davar Ardalan is the Founder and Storyteller in Chief of IVOW, a cultural storytelling startup powered by AI. Prior to this, she was a public broadcasting journalist for 25 years, most of those at NPR News, where she designed stories anchored in multiculturalism and steeped in historical context. In 2015, her last position at NPR was senior producer of the Identity and Culture Unit. Realizing that there is a gaping hole in AI algorithms that will define our future stories, she created IVOW, bringing together a team of journalists and technologists to design cultural IQ in AI. This includes leading narrative dataset challenges and AI storytelling projects throughout the world. Ardalan, who has also served as Managing Editor at Hanson Robotics, has been recognized with a 2017 NASA Team Leadership award for Space Apps, a Gracie Award from the American Women in Radio and Television and a shout-out in the popular comic strip Zippy. In May 2014, she was the recipient of a United States Ellis Island Medal of Honor, for individual achievement and for promoting cultural unity.View the profile
About the talk
What challenges and opportunities exist for quality leaders of tomorrow? Let's look at some emerging technology problems.
Today many AI products and tools are missing global voices and cultural data. This might mean that future AI systems will internalize the same lack of cultural diversity that we find in several media narratives -- unless there are interventions in collection, usage, and sharing of data about world cultures. Join Davar Ardalan, CEO of IVOW, an AI storytelling startup, as she explores ways to invite global voices into AI as we collaboratively build the foundations for Cultural IQ in machines.
00:08 QA testers will influence how citizens experience empathy in machines
00:10 Narratives in AI. Cultural engines
01:32 Startup IVOW and its idea
03:02 Why storytelling?
03:46 Why storytellers?
05:36 Problem and challenges of AI
06:38 How can stories help AI systems?
07:09 Cultural relevancy, citizen and their influence on AI
08:33 The age of AI DevOps can't neglect the critical roles of QA and testing
09:41 How do you expect to automate testing when it doesn't have your voice in it?
12:06 Future careers in QA and cultural IQ
The reason I want to start the conversation 00:03 with introducing the war is what we need. 00:07 We prepare the test leaders of today need to be prepared 00:12 to be the test leaders of tomorrow. 00:18 And I want to give Mike to the war right now and she will be talking about the new landscape. 00:22 And new challenges in new opportunities for us to be in this profession and what new rules probably will be emerging. 00:29 From those opportunities. 00:38
And then we're going to follow up with some discussions. 00:38 Of what it means for us and how we get ready for it so. 00:44 Start with the work. 00:48 Thank you so much. 00:48 It's such a great honor to be here. 00:52 Cortana we did meet in Ireland. 00:55 And. 00:58 It was an incredible meeting of the minds similar to hear my background is I'm a journalist, 00:58 I worked at NPR News in Washington for 22 years. 01:07 I've been the managing editor of Hanson Robotics working with Sophia, 01:11 the robot. 01:15
I'm also deputy director of the presidential innovation. 01:15 Membership program that's a tour of duty in American innovation, 01:19 which is coming to an end soon. 01:22 But it's a great honor to be here to talk to you about the future of AI and the lens that. 01:24 We are taking my startup is called "IVOW" and it stands for voices of wisdom. 01:32 We returned just a little bit ago from the United Nations. 01:39
We were invited to bring a delegation of global storytellers an AI experts. 01:43 The world is to have a full day of workshops, 01:48 interactive workshops on the future of storytelling an artificial intelligence. 01:51 And the idea that as children. 01:57 We learned by having our parents read books to our kids and. 02:03 wouldn't it be amazing if as we teach the future machines to be more socially intelligent and culturally intelligent. 02:11
That we also teach them about our stories. 02:19 Boyang Albert Lee is one of the most pioneering AI storytellers. 02:23 He is currently in San Francisco and he was Disney's first AI researcher. 02:28 So he talks about the fact that to function properly in a society. 02:34
Clearly, we, as humans rely a lot on. 02:39 the cultural awareness between each other as AI becomes more sophisticated it needs to have a deeper understanding of who 02:42 we are as humans and similarly, AI can play a major role in helping humans become more aware of different cultures around the world. 02:51
So why storytelling as I mentioned, 03:02 we grew up reading books and there's no reason why we can't do the same for machines and teach them about our history, 03:06 our communities, our myths and legends as I look around 03:16 I can tell that everyone here comes from a different amazing 03:19 cultural background, I don't know what that is, 03:23 but I feel like you guys would have so much fun if you spent. 03:26
Just 3 hours getting to know each other's backgrounds and where you grew up and who influenced you in your life was it your grandparents/ 03:30 What are the traditions that you are taking 4th and bringing to your children and the next generation. 03:38 So why storytellers? 03:46 Because storytellers are the keepers of culture in many parts of the world in the S Pacific in particular. 03:46 That's an oral storytelling community. 03:54
Not much is written until recently and storytellers are the ones who have been keepers 03:57 of culture and tradition. 04:04 Storytellers 04:04 can be the ones to usher in a new era of cultural IQ in artificial intelligence? 04:08 And storytellers can nurture a much deeper understanding of human cognition, 04:13 and in fact, there are amazing AI. 04:20
Researchers who come through the field of storytelling so it's all-around a comparative narratives 04:23 and conversational AI, but their research is about 04:30 computational narratives and how do you dissect a story and how do you understand the DNA of a story? 04:34 And how can you teach that to a computer. 04:42 These are the scientists and 04:42 AI experts who were with us in Geneva. 04:48 So Wolfgang Keys from the Crow Native American tribe. 04:51 He's an AI. 04:55
Here at Florida International University and he proved that the MIT story system can understand Crow literature, 04:55 so the MIT Genesis story system has been fed 05:04 primarily Shakespeare and Mark Twain to make it interesting and he said, 05:07 for his master's thesis, he wanted to prove whether 05:11 is culture could be understood by this machine and by only feeding it 100 pieces of data 05:14 he was able to do that. 05:20
Rafael Perez is from Mexico City and he's created a computer that generates narratives is called Mishika and 05:21 it generates folklore from the Aztec culture. 05:30 So the problem of course, 05:36 is that every day, we use apps to message our loved ones and games to battle are an imaginary force. 05:37 We use artificial intelligence to drive smarter decision making 05:44 the piece of the puzzle is missing an that's global voices. 05:49
The problem and the challenges are that AI systems struggle to be responsive for the most part to the values goals and principles of different communities. 05:53 And Mike 05:53 had mentioned around facial recognition. 06:02 This is something you've read about you know that these systems still are not 06:04 really able to be inclusive, 06:09 too many AI systems reflect the biases and perspectives of developers. 06:11
AI algorithms train on datasets to learn patterns, 06:16 which currently unlimited. 06:19 And understanding 06:19 global cultural contexts and the fact that the lack of cultural diversity and datasets will, in fact, limit the effectiveness of governments as they are creating smarter cities. 06:22 And businesses as they're looking to expand into new markets. 06:32 This is a quote from David Banks of Carnegie Mellon University. 06:38 He's an AI and ethics and policy scholar. 06:42
And he says stories provide a window into the ways people perceive interpret and value 06:45 the world. 06:53 And to move AI beyond the narrow needs and interests of the current developers 06:53 we need to fully engage 06:58 with the cultures and communities around the world and so he was one of the moderators at our summit in 06:59 Geneva. 07:06 Very soon cultural relevancy is going to need to be a two way 07:09 Street for AI solutions to speak to people globally. 07:14
Very soon citizens are going to 07:22 influence how AI services and solutions are being designed created and sold. 07:26 If you look at the growing Hispanic American population, 07:35 Google surveyed something like 5000 Hispanic Americans and they found that 70% 07:39 of the respondents said that it's important for the content. 07:46 That 07:46 they want to buy to be culturally relevant when they're gathering information about a purchase. 07:51 AI is only going to become... 07:51
Culture is going to be even more relevant in AI because if you're trying to create solutions for communities throughout the world. 08:02 And they download something and it doesn't culturally speak to them. 08:10
They're not going to use it, 08:14 which means that we're going to go into this constant cycle of creating 08:15 products that are not going to be effective globally that are only going to be effective selectively for the ones who are elite highly educated 08:19 and lived in a Western world. 08:29 The age of AI DevOps can't neglect the critical roles of QA and testing. 08:33
And I come from a journalism background and I look forward to the day when you are leading the creation of these solutions. 08:39 Along with storytellers and journalists. 08:48 Because you are the front lines. 08:50 A bringing empathy to machines. 08:53 QA testers will influence how citizens experience empathy in machines. 08:57 This can be something that inspires many people when you think about the Future, 09:03 and robots, and automation and 09:09 what place you can have in it? 09:11
I hear a lot. 09:14 I read a lot about how testing has to become 09:14 automated for it to be scalable and that manual testing potentially is not the way to go anymore. 09:20 I think that, emphatically, I reject that we are not there yet. 09:26 We have not created the infrastructure for this manual testing to go away and for everything to become 09:30 automated. 09:36 When the data that currently exists does not even represent you. 09:36 And your background. 09:41
So how do you expect to automate testing when it doesn't have your voice in it? 09:41 It's then an incredible opportunity for storytellers to work hand in hand to influence this new future. 09:51 Stephen Hawking in 2016, 09:51 said, that beneficial AI will take time. 10:01
And that's the category that this falls under is that who is willing to 10:04 speak up and say that we can't constantly be thinking about how to make money and not pausing and thinking about creating the infrastructures 10:10 that we need. 10:18 So one thing that we're working on is the idea of these cultural engines. 10:18 As you we have 2 big projects 10:26 which I'm excited that Anna is an advisor for one of them. 10:29
So we're going to be putting out 10:29 global data set challenge on the stories of women in history, 10:35 culture science and technology. 10:40 This is to create 10:40 algorithm that can capture the stories of women in particular. 10:43 The point here is that we're on a 10-year mission to bring cultural IQ 2 machines. 10:49
Initially, this data set will be used in our 10:49 cultural conversational AI named Cena, 11:01 but the data set is going to be public and it's going to be 11:03 published we hope on AI Commons. 11:06 AI Commons is a new international endeavor to bring datasets and methodologies around the future of AI. 11:06 2 portals where other people can look at these data sets, 11:17 and learn from them. 11:20
So imagine the future Wikipedia, 11:20 where you could go looking for a data set and you could go to AI Commons and read also about the methodology. 11:24 So this particular data set that we're doing would feed something like this, 11:30 where it's the data is being entered, 11:37 and various techniques will be used to attach features to it. 11:41
Clearly, they'll be annotations that are really important, 11:48 and then there's a critical role for testing because 11:52 we can't do this unless we make sure that what we're putting out there is also going to be 11:57 relevant and something authentic. 12:03 So a future career in QA and cultural IQ could be a cultural architect, cultural expressions leader. 12:06 This is incredibly daunting. 12:06 I mean, I'm not naive by the way I have 8 children. 12:18
My husband and I have 8 children. 12:21 We have raised 8 children, 12:24 the youngest is 20. 7 boys and a girl I understand what the future lies for them. 12:25 My granddaughter is 5 and when I went and visited her in Pittsburg. 12:25 She ran and she calls me Azis. 12:36 She's like Aziz, please come and meet my friend Alexa. 12:38 OK. 12:43 So guess watt. 12:43 Don't you want to be a cultural expression leader 12:43 for your granddaughter? 12:50
Buy this talk
Access to all the recordings of the event
Buy this video
With ConferenceCast.tv, you get access to our library of the world's best conference talks.