How GoFundMe combats identity risks while maintaining a human approach

Named by McKinsey as the fastest-growing type of financial crime in the US, synthetic identity occurs when fraudsters piece together authentic personally identifiable information (PII) to make a fake identity and use it to apply for loans and credit cards, or even start a fundraising campaign.

*Source: Alloy 2024 State of Fraud Benchmark Report **Source: TransUnion Analysis of Synthetic Fraud

Today’s all-time high levels are the result of many factors. High inflation rates, for example, might lead individuals to resort to fraudulent activities to generate income. New tools like generative AI make it much easier for fraudsters to generate synthetic identities.

Synthetic identity fraud’s risks are unique for crowdfunding platforms like GoFundMe, which exist to get funds for legitimate users in need. With trust paramount to recruiting and retaining users, stopping malicious activities such as fake beneficiaries and account takeovers (ATOs) is key.

“While synthetic identity fraud is a risk for any business, what remains in our control is how we respond to it,” said Matthew.

One of the hardest things about detecting synthetic identity for any business is just how sophisticated it is. Fraudsters often spend months building up a believable identity and credit history, bypassing standard ID checks. So it’s often only noticed when it’s too late.

As a result, there’s no single tool for identifying synthetic identity fraud. However, innovations in machine learning are making it possible for companies to look at more data points and build  more accurate risk scores at the onboarding stage, so their teams can investigate accordingly.

“Ultimately, it takes highly skilled teams performing critical tasks as well as tools that support their ongoing success,” said Casper. “As much as we want to remove bad activity from our platform, we’re very sensitive to keeping good activity,” said Matthew. “Over seven years, we’ve built out mutually complementary processes and technologies that enhance the safety of our platforms. And that’s helped us retain customers in significant volumes that otherwise we may not have been able to support.”

With the ultimate goal of keeping funds going to legitimate users in need, Adyen and GoFundMe have developed a robust approach to fraud prevention that makes tackling the challenge of synthetic identity a highly collaborative process. Here are some examples of our strategies.

1) Strong KYC and identity verification A waterfall approach to KYC and identity verification with strong providers and backup coverage is the first port of call for GoFundMe and Adyen. Tactics like liveness and selfie checks can be used to help combat potential identity fraud at this stage, while staggered onboarding helps reduce the risk of drop-off. 

“In addition to obvious table stakes of traditional KYC and identity verification like name, date of birth, address, etc.,” said Matthew, “there’s activity. To ignore that adjacent information is to miss so much context and a greater ability to have a safer, more dynamic business. And so we’ve really started looking at identity as just one part of a holistic person online — yes, it’s who they are, but also what to expect from them when they come to use our platform.”

2) Data linking and risk scoring Using tools such as Score to link data and intercept fraud rings is just one example of how GoFundMe uses Adyen for Platforms to prevent abuse. Since Score provides insights via a broad set of risk signals, it also helps reduce the platform’s time spent on user security reviews. In addition, GoFundMe continuously updates blocklists and assigns risk scores to users.

“If someone is committing fraud without knowing it, or if our computer says no, we don’t just want to shut down personal needs,” said Casper. “I think Adyen and GoFundMe have been able to do that very well with an automated approach paired with teams working in the background to make it happen.” 3) Enriching assessments with mutual data sharing In some cases, a second assessment by the other party can strengthen an assessment and lead to different conclusions. ATOs, for example, often involve fraudsters using synthetic identity information. Maintaining a list of data points used in ATO scenarios helps GoFundMe avoid offenses, block bad actors, and minimize impacts on campaign owners.

“Together with data from our entire platform, we're able to enrich the data we get from all the different sides and different platforms,” said Casper. “It's really helpful and GoFundMe’s teams can take action on those scores and decide whether they want to blocklist someone or continue the relationship based on the information provided.”

4) A human and collaborative approach to operations Last but not least, refining processes where experts review issues as they’re flagged by technology or reported by users makes it possible to ensure funds keep flowing to legitimate users. There are countless examples where automated tools resulted in a “computer says no” situation, but Adyen and GoFundMe collaborated closely to improve a decision with data. 

“For example, a couple of years ago, we had a fundraiser with eight or nine transactions on the same card, all different names,” said Matthew. “That sets off alarm bells on our end: what's happening here? And then, once we actually started reviewing things, it turned out that the donations were to a fundraiser for a dog, and each of the names was the cardholder's dogs. What a good moment to educate the donor and prove the impact of our human approach.”

While factors like high inflation rates and artificial intelligence have created a fertile ground for synthetic identity schemes, Adyen and GoFundMe have found that a balanced, human approach is essential. Automation offers efficiency, but collaboration between platforms and payment service providers (PSPs) is crucial to safeguarding legitimate users.

Leveraging innovative tools like Score and KYC advancements can enhance risk management by enriching data and enabling swift responses to fraudulent activity. However, challenges persist in detecting sophisticated fraud while ensuring a seamless onboarding process. Robust KYC verification, advanced fraud detection tools, and collaborative efforts, are vital solutions in navigating these challenges effectively.