Trust and Safety Architect Jane Lee on Misinformation, Investigative Thinking, and Making an Impact

Jane Lee, Sift’s newest Trust and Safety Architect, knows a thing or two about fighting fraudsters on the ground. Once a legal case assistant-turned-private investigator focused on insurance fraud, she’s since built an impressive career in online trust and safety at some of the world’s largest, most innovative tech companies.

While her razor-sharp, critical-thinking skills have given her the ability to deftly analyze and assess risk, it’s her commitment to real-world impact that’s truly shaped the way she works, and why. We sat down with Jane to discuss her professional path and how an eye for anomalies has allowed her to help businesses protect people and platforms.  

How did you end up in the trust and safety space?

I most recently came from Facebook, where I worked as a project manager on the Site Integrity Operations team—the team that works on spam prevention. Prior to that I was at Square, where I worked on chargebacks and got my first taste of payment fraud. But to explain how I ended up in trust and safety, I have to rewind to my first job out of college, where I worked at a law firm as a case assistant. I somehow fell into taking care of billing discrepancies—it gets really complicated, and I was given the responsibility of figuring out who owed what money to whom. Following that job, I joined a tiny firm as a private investigator; I investigated worker’s compensation fraud, or insurance fraud, and that’s where I think I got my big investigative chops and really honed my critical thinking skills. 

Insurance companies would send over claims they suspected of being fraudulent. I wasn’t the one doing surveillance—there was a separate team for that—but I reviewed the flagged claims and recorded any inconsistencies that I found. I never gave opinions or recommendations, but provided more of a scientific report of my findings.

After that, I got my big break in trust and safety in tech. I joined the Disputes team at Square, and was able to combine my billing experience and my P.I. experience to investigate chargebacks. From there I went to Facebook, where spam is a trust and safety issue. I was there for five years, and for the first two years I was there, most of our spam response was very reactive and based on user reports, where we would manually assess whether content violated our policies and qualified as spam. Down the road we automated a lot of that, and shifted to a more proactive approach based on patterns and trends we saw coming out of the data. 

With U.S. elections on the horizon—and “fake news” making national headlines—can you share some insights from your time working with misinformation in global elections?

When misinformation claims and issues spiked after the 2016 election, my company mobilized a team to handle it—with the scale of problems and the size of our user base, we wanted to invest in resources for things that truly mattered, that were worth tackling. My expertise is in harmful landing pages—basically, malicious websites, what they do, and how bad behavior can manifest on different domains. So, I was tasked with figuring out how to operationalize the detection of misinformation at scale, as well as determine if there were any enforcement policies we could put in place to identify it. For example, were there certain backend signals contained in a subset of misinformation websites or pages that we knew to be spam, that we could then use to determine spam patterns or features?

I define spam as unsolicited monetization. So, we worked to identify groups of misinformation websites acting as part of connected networks, that were using ads and other content—click traffic, basically—to monetize their scams. We were able to identify small but viral networks, or fraud rings, that were participating in this kind of behavior and that we were able to crack down on. It wasn’t just posts and comments on our site, either, but user accounts that were associated with those networks, and other user-generated content. 

You have all of these different layers of fraud that can occur, so it gets very complicated. Add in the fact that fraudsters will work across different connected networks, and even websites, to execute spam attacks, and you can essentially guarantee that if a certain type of fraud is taking place on, say, one food delivery app, then it’s also taking place on its competitors’ apps. And that’s one of the most difficult things about dealing with scams, spam, or any type of content fraud—it’s widespread, hard to detect, and from a user standpoint, pretty easy to fall for if you’re not actively looking for it.

Let’s talk about COVID-19 and its impact on e-commerce. What are trust and safety teams facing?

COVID-19 came along and just disrupted e-commerce in every possible way, and trust and safety teams are facing a lack of access to accurate historical data. The data they do have can’t necessarily give them the predictability they’re accustomed to. And, whenever there’s any kind of global current event like this, it will be exploited by scammers and bad actors; we know that  fraudsters are trying to capitalize on the crisis through things like stimulus check scams and vaccine scams. Friendly fraud is on the rise, too, likely because people are short on cash. Chargeback disputes are increasing for a lot of businesses, and filing the paperwork to fight them takes up a lot of time. All of that falls under the trust and safety umbrella.

But one of the biggest challenges these teams are dealing with is happening at companies that have had to reduce their workforces. In my experience, operational roles are often the first to be scaled back when companies have to go through any type of reorg, and it leaves those teams very lean. All of that mounting fraud and the increase in chargeback disputes are being handled by fewer people, and that doesn’t only create slowdowns—it means less time for manual review, way higher likelihood of decision fatigue, and more opportunities for human error. I don’t think we’ll truly know how big the overall impact of coronavirus and the recession really is until a year or two years from now. Probably longer.

A quick note about those rising chargebacks: I’d advise trust and safety teams to really focus on the ones they typically have higher win rates for. When you’re faced with an onslaught of new cases and smaller teams to handle them, it’s usually more beneficial to have spent your time and resources on safer bets.

What brought you to Sift?

I’ve been following Sift for awhile. What excites me most is that there’s such a huge opportunity to have cross-industry impact. I would love it if companies could communicate with each other more to solve their fraud problems, instead of each organization being siloed and attempting to tackle this big problem on their own. Because Sift has a huge global network, this massive client base, and this really valuable knowledge and data that can truly make a dent—an effective dent—in these fraud networks, we have a real chance to move businesses past the constant game of cat-and-mouse they’re stuck playing with fraudsters. 

Connect with Jane on LinkedIn.

***

Download Sift’s new report, Digital Trust & Safety Index: Content Abuse and the Fraud Economy.