The phenomenon of GoFundMe has allowed people with compelling stories, or who are just down on their luck, to raise money. This morning, I woke up to the (horrifying) realization that a person who has been harassing someone (*cough*) on the internet is now crowdfunding to pay for it! It got me thinking, what is the data behind fraud in crowd funding, and what can we do about it?
The following may come as a total surprise, but it turns out that not everyone tells the truth on the Internet. Some people might even say things that are not true, because it's in their financial interest. I know, it's shocking and heartbreaking. I'm aghast also.
As one academic paper put it:
Unsurprisingly, deception and fraud in such platforms have also increased, but have not been thoroughly studied to understand what characteristics can expose such behavior and allow its automatic detection and blocking. Indeed, crowdfunding platforms are the only ones typically performing oversight for the campaigns launched in each service.1
Not content to simply let good scams go to waste, academics have sent out to use advanced analytics and machine learning to identify scams, and train their algorithms on “scam versus not scam.”
we propose to use the feature selection method called Forward Stepwise Logistic Regression… our model classifies scams and non-scams with 87.3% accuracy. We found that features extracted from the Comments (6 out of 17) section are most helpful in detecting scams, closely followed by those from the campaign creator’s information (4 out of 17) and the Updates section (5), then by the Campaign section (2).2
As in most things on the Internet, the comment section is especially rich data! Not content with only an 87% accuracy, subsequent researchers designed ever more complex models to identify scams, with remarkable improvements and accuracy thus far, achieved largely through including the images used to advertise the campaign, and additional data that can be had at day zero— before comments give it away:
we build machine learning classifiers, and show that it is possible to automatically classify such fraudulent behavior with up to 90.14% accuracy and 96.01% AUC, only using features available from the campaign’s description at the moment of publication (i.e., with no user or money activity), making our method applicable for real-time operation on a user browser.
Yes, that paper is actually titled: “I Call BS: Fraud detection in crowdfunding campaigns” That's not a joke. “I call BS” is the actual name of what might be my favorite title of an academic paper of all time.
To be fair, globally, a tremendous amount of crowd funding goes to paying for medical expenses, but even this is not as virtuous as it seems. One paper from China evaluated the socioeconomic variability in crowd funding for specifically medical expenses, and found in the direction of those already wealthier:
The study finds significant differences in crowdfunding outcomes among different socioeconomic groups. Higher SES brings higher donation amounts, higher proportions of fundraising targets reached, a higher number of donations received, and more shares on social media. This socioeconomic gradient in crowdfunding outcomes contradicts the social expectation of prioritizing help to the neediest in medical crowdfunding.3
I might take issue with the Andrew Tate-ism that “your network is your net worth,” it may be the case that your network is your stealth health insurance policy, if this research from China is generalizable. The most common kinds of medical fraud using crowd funding have been categorized by yet more academics:
We identified the following categories of fraud: faking/exaggerating one's own illness (n = 29), faking/exaggerating another's illness (n = 7), impersonation (n = 8), and misapplication of funds (n = 8).4
Faking somebody else's illness? That's sometimes called Munchausen's by proxy, and we've covered it in the newsletter before.
What to do about such scams? It's tempting to think the same technology that created the problem can solve the problem, but in what is now my second most favorite article title of all time, authors point out that this is not the case…
We Keep Us Safe, Venmo Doesn't: The Impact of Peer-to-Peer Payment Apps on Mutual Aid and Community Organizing
Automated solutions to crowdfunding fraud often penalize those who are the most vulnerable:
However, banking regulations, in-app policies, and tax legislation-which prioritize fraud prevention, minimize platform liability, maximize profits, and tax organizers and mutual aid recipients-can hinder mutual aid and community organizing.5
The individuals who need the aid the most, and who can least afford to have their account frozen? They are often the least wealthy and least privileged am among us. They are more likely to be women, and they're more likely to be people of color. It's not as simple as just using technology to freeze these accounts. It can really leave desperate people in the lurch.
There's no free lunch when it comes to preventing fraud online, we may have to use some good sense. A little bit of skepticism and may be in order. For example, if somebody is raising money for legal fees, those cases are filed publicly. You can look it up, and see what they're about.
One example of this is defamation, in which people may easily claim they're being sued, and it's expensive. I've seen people raise a lot of money on the claim that they “are being sued.”
I'm obviously not going to draw attention to those individuals. However, it's worth noting that most defamation cases against private citizens are often represented by attorneys retained under the home insurance policy of the individual. Yes, home insurance policy often cover legal defense when being sued for defamation. Interestingly, at the end of the case, if it's that found the person did willfully committed defamation, the policy can decide that the defamer is on the hook not just for the judgment, but for the cost of the defense, too.
We need to be a little bit more skeptical, people. When it comes to medical fundraising, it might make more sense to change the perverse incentives of a medical system that sucks Trillions out of the economy in the US every year. It’s at the expense of the sickest among us. I happen to think canceling Medical debt and advocating for financial hardships in the care of individuals who faced those bills is a better strategy than raising tons of money… just to pay it back to organizations that make more money than any of us can imagine.
Thanks for reading, and have a remarkable day. Don't get scammed!
Perez, B., Machado, S., Andrews, J., & Kourtellis, N. (2022, June). I Call BS: Fraud detection in crowdfunding campaigns. In Proceedings of the 14th ACM Web Science Conference 2022 (pp. 1-11).
Lee, S., Shafqat, W., & Kim, H. C. (2022). Backers Beware: Characteristics and Detection of Fraudulent Crowdfunding Campaigns. Sensors, 22(19), 7677.
Cheng, C., & Ren, Y. (2024). The paradox of seeking help: Socioeconomic inequality in medical crowdfunding in the era of the mobile internet. Chinese Journal of Sociology, 10(1), 100-130.
Zenone, M., & Snyder, J. (2019). Fraud in medical crowdfunding: A typology of publicized cases and policy recommendations. Policy & Internet, 11(2), 215-234.
Rankins, M. G. (2023). We Keep Us Safe, Venmo Doesn't: The Impact of Peer-to-Peer Payment Apps on Mutual Aid and Community Organizing. NC Bank. Inst., 27, 240.
A lawsuit is not harassment. G-d knows what you did and He will in due time bring you to justice.