I saw the headline and cringed. But not because I was worried about snooping landlords.
That story, in the Washington Post, featured a startup bragging that it could turn social media profiles into reports for landlords. The reports would rank applicants’ personality traits and financial stress levels, and would even note if they mentioned a pregnancy. “If you’re living a normal life, then, frankly, you have nothing to worry about,” quipped the company’s co-founder.
If you read tech journalism, you may be convinced that your Facebook posts, who you follow on Twitter, or the previous employers you list on LinkedIn are being constantly laundered through secret algorithms to make scores that creditors, landlords, and employers use to judge you.
You may have heard, for example, of the fabled “Facebook credit score”:
These headlines probably get lots of clicks. But the picture they paint is just plain wrong.
I work at Upturn, where we’re constantly thinking about how technology shapes opportunity and inequality. (This year, for example, our team helped advocates work with Google to ban payday loan ads and rated police body camera policies on civil rights.) I used to enforce consumer protection laws, and, before that, I was a privacy advocate.
The idea that a person could be automatically denied credit or housing based on their social media footprint is exactly the kind of “big data” redlining that I try to guard against.
There’s just one problem: Social media scores — scores designed to be used for important decisions beyond marketing — are pretty much a myth. Facebook is not giving out credit scores based on its users’ profiles and friends, and it doesn’t want anyone else to, either. Any company that claims to be doing so is probably full of it.
Why do I say that? We’ll get there. But first, here’s why we should care:
This hype distracts from real problems.
I see it firsthand. I sit on panels about access to credit that get derailed by rumors of these “secret scores.” I field questions from puzzled journalists. I read concerned e-mails and Facebook posts from family and friends (“is this real?!”). (No. It’s not real.)
In the meantime, important issues have to compete for attention. Three examples: The Consumer Financial Protection Bureau says 35 million people in the United States — many of them poor and minorities — face substantially reduced access to credit because they lack a credit record sufficient to produce a credit score. The laws that govern debt collection — a frequently abusive practice — haven’t been revamped in 40 years. And there are still far too few safe, short-term loans for people who need them.
I want everyone who cares about inequality in the digital age to think more about real, unsolved problems, and less about the smoke-and-mirrors world of big data scores.
. . .
In that spirit, here’s a rundown of why social media scores aren’t a big deal today, and won’t likely become one in the near future:
1. Social networks have started to prohibit scoring companies from accessing their data.
Earlier this year, Facebook adopted a new rule for outside developers:
Don’t use data obtained from Facebook to make decisions about eligibility, including whether to approve or reject an application or how much interest to charge on a loan.
The language speaks for itself. Any company claiming to use Facebook data to make serious decisions about people is out of line.
These rules make good sense, and I expect we’ll see more of them. Today, at least, social networks are focused on making money from ads. They want their users sharing, not looking over their shoulders. And they probably don’t want to risk triggering new legal obligations. Which leads me to…
2. Federal laws in the United States
Technology is outrunning many laws. But the laws we do have, on this front, do have some power. Here are two important ways they are working:
Social networks don’t want to look like credit bureaus, which are actually regulated. The Fair Credit Reporting Act requires a wide range of companies that collect data about people for the purpose of certain “eligibility” decisions (including lending, employment and insurance) to meet a raft of legal obligations, including accuracy of data. Today’s social networks don’t want to risk having to take on these responsibilities.
Lenders, employers, and landlords could get in legal trouble for using social media scores. A range of federal laws, including the Equal Credit Opportunity Act, Title VII of the Civil Rights Act, and the Fair Housing Act can trigger liability when seemingly neutral practices cause lead to unfair results (“disparate impact”).
Today’s mainstream credit scores (e.g., the famous FICO score) have been carefully tested, and are sold with the guarantee that they are legally compliant. Social media scores won’t be able to offer same guarantees for quite some time (if ever).
3. We don’t know if social media scores are actually good at predicting anything.
Even in the relatively well-studied field of credit, the predictiveness of social media-based scores is ambiguous at best. When traditional credit data is available (like how you pay your bills), social media data adds very little value.
. . .
In closing, a couple of caveats.
Companies outside of the United States are starting to make important decisions about people based on their Internet behavior. These experiments are being conducted under different legal regimes, and typically try to address the absence of a robust consumer reporting infrastructure. They deserve scrutiny.
And regardless of where you live, it’s reasonable to worry that an employer or landlord might try to peek at your social media accounts, even though they probably shouldn’t. So, yes, check your privacy settings.
Just don’t get too distracted by the social media score hype. This isn’t an issue yet. We have real problems to solve.