The nomination of Amy Coney Barrett to the Supreme Court, and the related existential threat to the Affordable Care Act, got me to thinking about something even more troubling: the many ways in which society takes information about people - often on qualities or events that are irrelevant or out of their control - and uses it against them.

It's a tendency that we need to curb.

The ACA is a notable case where Congress did something to restrict how personal information can be used. The law forbids health insurers to deny coverage or charge extra to people with pre-existing conditions - or even to ask customers about their medical histories. This makes sense, because if the companies can avoid anyone who might get sick, insurance ceases to be insurance, and a single illness can become the beginning of an inexorable spiral into penury.

Politicians on both sides of the aisle seem to recognize the importance of the pre-existing condition clause. Ted Cruz recently supported keeping it even if the Supreme Court strikes down the ACA, and President Donald Trump has promised to do so through an executive order (not that his word means much). Tens of millions of Americans have pre-existing conditions, and they'll rightly blame Republicans if they get dinged or priced out of insurance.

Newsletter signup for email alerts

This is particularly crucial in the time of covid-19. Infections have added a pre-existing condition to the records of millions of people, many of whom have also lost their jobs and will be seeking health insurance in the coming months. If insurers can demand medical histories - or get information on infections from tracing apps, which are vulnerable to hacking and aren't necessarily covered by medical privacy laws - the pandemic will vastly expand the ranks of the dis-insured.

So limiting the information that companies can see and use has huge benefits in health care. But why stop there? Other types of pre-existing conditions - the neighborhood where a person lives, a person's network of friends - are regularly used to make decisions such as who gets a job interview, who gets an apartment, who gets a loan and who gets granted parole. All too often, the information is biased in ways irrelevant to the decision being made, or serves as a proxy for protected characteristics such as race, age or gender.

I'm not suggesting banning the use of valuable background information. Rather, the kind, scope and age of information allowable should be carefully circumscribed, keeping in mind what is appropriate for a given situation.

Consider the millions of Americans, disproportionately Black and Hispanic, who will suffer eviction during the coronavirus crisis. What might have been a temporary setback, completely beyond their control, could end up haunting them for years, as third-party data brokers report it to potential landlords. A "pre-existing conditions" rule in housing could at the very least remove such events from a person's record after a designated period of time - as credit reporting bureaus already do with late payments for the purpose of computing credit scores.

Restrictions should also apply to using data to infer information about people. In "Ban the Box" states that forbid asking about a job applicant's criminal record, for example, employers can infer the likelihood of convictions from other indicators, such as Zip codes and gaps in consumer histories. This is similar to what Facebook does to target ads, classifying people primarily by demography instead of understanding specific qualifications. It makes lucky people luckier while denying opportunities to people born in the wrong place. In a nutshell, it short-circuits the American Dream.

A more generalized approach to "pre-existing conditions" would help prevent irrelevant or old information from being unfairly used against us, just as the ACA's pre-existing condition clause prevents insurance companies from discriminating against the most vulnerable among us.

Cathy O'Neil is a Bloomberg Opinion columnist.