Trust and Fraud Detection at Scale: Instagram’s Stephanie Moyerman

**Sam Ransbotham** (0:02)
The emerging field of data science is progressing rapidly, resulting in numerous managerial challenges.
How do technology leaders ensure quickly evolving digital spaces like social media stay safe and equitable? Find out on today's episode.

**Stephanie Moyerman** (0:17)
I'm Stephanie Moyerman from Instagram, and you're listening to Me, Myself, and AI.

**Sam Ransbotham** (0:22)
Welcome to Me, Myself, and AI, a podcast on artificial intelligence and business. Each episode, we introduce you to someone innovating with AI.
I'm Sam Ransbotham, professor of analytics at Boston College. I'm also the AI and business strategy guest editor at MIT Sloan Management Review.

**Shervin Khodabandeh** (0:40)
And I'm Shervin Khodabandeh, senior partner with BCG and one of the leaders of our AI business. Together, MIT SMR and BCG have been researching and publishing on AI since 2017, interviewing hundreds of practitioners and surveying thousands of companies on what it takes to build and to deploy and scale AI capabilities and really transform the way organizations operate.

**Sam Ransbotham** (1:06)
Stephanie, thanks for joining us. Tell us a little bit about your job. What do you do right now?

**Stephanie Moyerman** (1:11)
I am the data science director supporting Instagram wellbeing. Instagram wellbeing ensures trust, safety, and integrity on Instagram's platform.

**Sam Ransbotham** (1:20)
We've heard a lot, I guess, about the various dark sides of social media that seem like there's a story almost every day about something that's gone wrong. I think we overlook all the things that are going right.
What kind of role does artificial intelligence and data have in helping to bring the light back to social media?

**Stephanie Moyerman** (1:38)
There's a wealth of data that actually show that across the sweeping industry and not just social media, but across a general industry for teens, safety over the last 100 years, wellbeing over the last 50 years has actually increased for teens.
And there's an interesting data story on that if you look at everything from what you think of as your most severe potential problems, such as police reporting, to just the ability to access information, to teach yourself, and really as simple as literacy rates have just increased time over time again. And it's really the data that gives us the lens on that. So being able to analyze the data on a platform like Instagram, one of these broad social media platforms where you have access to tens of millions of teens and say, hey, this is how we can actually tell a story that makes this platform safer, better, and makes the world a slightly better place, is something that previous to the last 20 years, we actually just could not do at all.

**Sam Ransbotham** (2:39)
Yeah, that's a great point. I mean, we just didn't have that data at all in any way. There's no collection of it at all. Tell us a little bit about what you learned about how things are progressing.
You said, literacy rates. What about other things?

**Stephanie Moyerman** (2:50)
Within our own ecosystems, where we see things progressing is in the ability to help set healthy social norms. And this is how we think about things, not just for teens, but for the platform all up, how do we encourage and foster relationships amongst people and a course of dialogue that helps establish a normalized behavior that increases awareness of the good that's going on and sets a tenor for community conversation. And all of this is driven by large-scale personalization models. Models that try to predict what it is that will elicit the appropriate responses to a given situation.
On the opposite side, the entire industry actually, across e-commerce, across social media, has large-scale models of AI that have enabled us to understand the sentiment of conversation, the safety of conversation, the content within images that protect people from seeing things that otherwise would have gone unregulated. So if you think about sort of tabloid publications or things anyone could throw at your door before that were largely unregulated, now we actually have not only the breadth to do this, but the sweeping AI tools that allow us to do it at scale.

**Sam Ransbotham** (4:04)
That's beautiful because you're pointing out, how so much of this stuff just was happening beforehand, and now we recognize that some of it's just a measurement and an observation problem.

**Shervin Khodabandeh** (4:14)
That's true. It also feels to me that there's two factors going on, and maybe there's a race here because there is, on one hand, more and more people have access, and this is a sign of a free and thriving society, and people should have freedoms to post things that they want or share things they want.
On the other hand, the chances of bad or hurtful content has increased because they're just more participants. So that's one factor going on.