Opinion: Is Your Credit Score Becoming a Real-Life ‘Black Mirror’?

Fans of the dystopian TV series “Black Mirror” may remember the episode “Nosedive,” in which everyone in society is subject to a social scoring system based on daily interactions.

You rate others, they rate you, and your resulting “score” helps determine things like where you live, where you work, and who you associate with. (Spoiler alert: It doesn’t go well.)

That’s obviously fiction. But your credit score — the latest models of which are ingesting more and more bits of information, and attempting to be more predictive than ever — might increasingly bring “Black Mirror” to mind. 

Credit scores don’t just consider traditional elements like charges you’re making on credit cards, and whether and when you are paying them back. Instead they are assembling a much fuller financial picture, one that includes factors like rent payments or utility bills.

Take the recently introduced UltraFico model, which uses “cash flow insights” — how well you’re managing spending and whether you’re maintaining positive balances — to come up with a more comprehensive credit profile.

You agree to link your financial accounts, share your data, and FICO claims this will be “additive” to your score, boosting it rather than dragging it down.

The VantageScore 4.0 model also broadens the type of information fed into scores. It uses machine learning and trended credit data to glean more insight about whether your finances are headed in a good or bad direction.

With influential organizations like Fannie Mae now relying on these newer models, the trend looks to only increase from here.

It’s obvious why financial institutions would be on board with this trend. They desire as clear an image as possible of one’s financial behaviors before extending credit because that helps them make better lending decisions.

On a positive note, it could also result in increased credit access for millions — particularly lower-income consumers — since under traditional scoring models, many Americans have gaps in their file that make lenders reluctant to do business.

For example, 18% of U.S. adults have limited or no credit history, even though almost everyone (96%) has a bank account, according to FICO.

How Much Information Do You Really Want to Share?

Fair enough. But at what point does the shared information become so voluminous and granular, that it becomes something akin to surveillance?

After all, not all financial behaviors are positive. Many of them are distinctly negative, and that might be a level of scrutiny you’re not comfortable with. If you let credit scoring agencies peer over your shoulder like that, day in and day out, could their judgment potentially put you in a hole you can’t get out of?

That seems like a devil’s bargain. If the cost of credit access is essentially constant behavioral tracking, that’s a heavy price to pay.

Consider the real-life case of China’s social credit system, which was rolled out in some pilot projects around the country to encourage positive behaviors (and punish negative ones). Reducing people to a number may appeal to governmental planners aiming for more societal control, but there’s no doubt it’s a slippery slope.

When does the flow of shared information become so extensive and detailed that it crosses the line into surveillance?

Look at what happens in “Black Mirror”: The main character’s score starts to slide, and negative interactions compound until her life is essentially ruined, and it’s almost impossible for her to claw her way back.

Is that scenario really so far-fetched? If some Artificial Intelligence program scans your financial behaviors and decides you’re not worth the risk, that can have a very real impact on your life, such as being denied for a car loan or a mortgage or a job opportunity. And what’s your recourse then — to complain to a machine?

An AI Backlash is Brewing

It seems that scoring agencies are anticipating some Big Brother-related criticisms here. FICO, for instance, is quick to point out that these new scoring mechanisms are “consumer-permissioned.” If you don’t want to share all this additional information, then you don’t have to.

As it should be. But it also begs the question — will it always stay that way? Or might there come a time when sharing that information is not so voluntary?

To be honest, this trend toward Too Much Information may affect lenders’ bottom lines, as well. Become too finicky about who you’re offering credit to and that slaps a hard cap on your growth prospects since many Americans’ finances are not in a great place right now.

On the upside, new credit scoring models could also expand credit access for millions of Americans, especially lower-income consumers whose traditional credit files are too thin or incomplete to give lenders confidence.

There’s also the broader question about whether this predictive push is even accurate. The underlying assumption is that the information AI is generating is always true, but that is just not the case.

AI sometimes hallucinates, or invents facts — often in an attempt to please whoever is asking the question, rather than admitting the limits of its knowledge.

That brings to mind another onscreen parable, Tom Cruise’s “Minority Report,” where cops try to predict and prevent events that haven’t happened yet. Try to see into the future, and you might be right — or you might not. 

That’s the riddle at the heart of the lending business, and it’s hard to blame financial firms for trying to amass as much data as they possibly can. But there’s a limit to how much personal information should be shared, with companies generating a score that will affect your daily life in multiple different ways. 

We’re getting close to that line, if we haven’t crossed it already.