We call on the field to recognize that applications that aim to believably mimic humans bring risk of extreme harms. Work on synthetic human behavior is a bright line in ethical Al development, where downstream effects need to be understood and modeled in order to block foreseeable harm to society and different social groups.
Fortune had it that when George Ball was looking to get a dehumidifier for his Indiana home, the house had one sitting in the basement from the previous owners. So he took it up to the bedroom, plugged it in, and went downstairs for dinner. A while later, he went back upstairs and found that the dehumidifier was on fire, with flames reaching from floor to ceiling. He called 911, and he and his wife, luckily, were able to put the blaze out.
Ball was perplexed when the firefighters who responded to his call pulled the device onto the front lawn and started taking pictures of it. They recognized it because it had been recalled years earlier, for being a fire hazard. “Oh my gosh, it shouldn’t even be in your house,” Ball recalls the fire chief telling him. “I almost burned our house down,” he says. “I had no idea.”
Adding a layer of irony to Ball’s case is that studying product recalls is literally what he does for a living — he’s a professor of operations and decision technologies at the Indiana University Kelley School of Business who specializes in the matter. “I spend my whole life studying this stuff, and it almost killed us,” he says.
Product recalls happen when a manufacturer asks consumers to return a product because the company has discovered that product is dangerous or has a defect that could render it unsafe. It’s not the sexiest of issues, but it is an important one: It’s not ideal when something sitting inside your home could hurt you or your family. And the United States’ recall system is a flawed one, perhaps especially on consumer products, which we’ll focus on here.
According to a recent report from the US Public Interest Research Group (PIRG), a watchdog group, the Consumer Product Safety Commission (CPSC), the government agency that oversees consumer product safety in the US, announced 292 product recalls in 2022, a 33 percent increase from the year before.
Recalls often take months or years to be announced, well after companies become aware of a problem. And recalls are sometimes announced multiple times, an indicator that injuries continue to occur even after the initial recall — and a sign of just how hard it is to get the word out to people that items in their households may put them in danger. You may have heard of some high-profile recalls, like the Ikea dressers that tip over or the Fisher-Price sleepers that kill babies, but not hundreds of others.
“Our recall system itself is broken because there just aren’t good ways for consumers to find out,” said Teresa Murray, consumer watchdog at the US PIRG, in an interview. Companies spend millions and millions of dollars telling us how great their products are — they’re not so eager to let us know when there’s a problem. Underlying the questions of when recalls happen and how consumers find out about them is a much bigger issue: how these items get to market in the first place. “The problem is not that products get recalled,” Murray said. “It’s that they’re not safe to begin with.”
There are generally two reasons for a product to be recalled — it’s designed incorrectly, or it’s manufactured incorrectly. Companies are constantly pushing on innovation and cutting costs, which is where issues of defects can spring from.
“If you’re a super cost-conscious product like a toy or something that is not highly innovative, and you’re trying to sell a lot very cheaply, when you go down in cost, you’re often up against cheaper suppliers, cheaper materials, cheaper labor — those things will eventually build up and you’ll make a mistake on the production line. That’s on the manufacturing side,” Ball said.
On the design side, if a product is really innovative, companies might not know what is going to happen until it gets to market, and that’s when they see failures come in. Often, the manufacturers are aware of some risks, but until they’re observed in real life, they’re more like statistical chances, the risk of a thing that might happen. Other times, there’s just no concept that there’s a problem until complaints start coming in.
Once issues do start to arise and companies become aware there’s something wrong, they’re not exactly eager to publicize what’s going on. At-home exercise company Pelotondragged its feet on alerting regulators and the public about incidents of children, pets, and objects being pulled under its treadmill. And once the CPSC called for the equipment to be recalled, Peloton initially resisted, saying its claims about the dangers of the product were “inaccurate and misleading.” It eventually changed course and did recall the treadmills, and the CPSC fined it $19 million for the delay in reporting the issue, one of the largest civil penalties in the agency’s history.
But for the most part, aside from having a moral compass, companies just don’t have much incentive to recall products, or to be loud about those recalls when they do. They don’t want damage done to their brands. Just how severe that damage may be depends on the potential impact to consumers, said Ross Steinman, a psychology professor at Widener University who focuses on consumer behavior. It’s one thing if the defect kills you, it’s another thing if it doesn’t meet a small regulatory standard that doesn’t really affect people’s lives. Even then, the reputational risk is often temporary. “A number of consumers report that they would not use the brand short-term, and that might be the case, it does depend on the magnitude,” Steinman said. “We’re creatures of habit, and if a brand has strong equity, in all likelihood, consumers will return to that brand.”
Whether to recall a product can be a hard decision for companies to make, internally. Research shows that in the consumer product industry, companies are likelier to issue a recall under a new CEO, especially when the prior CEO was forced out — it’s easier to place blame on them. As a CEO’s tenure goes on, companies recall less.
Ball’s research on the medical device industry has found that women leaders are, on average, more open to recalling faster than male leaders and that the more shares a CEO owns of a firm, the slower they are to recall, especially when problems are very serious. It’s quite possible that extends to the consumer product industry — except the problem is that there isn’t a ton of good data coming out of the CPSC. Compared to the FDA, which oversees drugs and medical devices, and the National Highway Traffic Safety Administration (NHTSA), which handles transportation, the CPSC falls short in a number of ways.
For one thing, the CPSC oversees an incredibly diverse set of some 15,000 kinds of products, from baby strollers to blinds to ovens. It’s impossible for them to keep tabs on them all, especially given that, like many government agencies, they are strapped for resources. Ball said it also “has relatively poor recall and complaint data compared to NHTSA and FDA, does not conduct inspections like FDA or mandate recalls like NHTSA, and seems to have little influence over consumer product firms unless or until it becomes a major crisis,” like with Peloton. “Companies are able to wait longer than they should or hide things or play games more than the medical and auto industry are,” he said.
A CPSC spokesperson said that the agency has recall data “sufficient to do our job” and it cannot comment on other agencies. The spokesperson said recalls happen fastest “when companies self-report and initiate the recalls themselves” and that most recalls are voluntary. “When firms refuse to enter into a voluntary recall, CPSC is focused first on alerting consumers to the hazard and urging them to immediately stop using and dispose of unsafe and dangerous products,” the spokesperson said.
It might be more helpful, say, if the CPSC, FDA, and NHTSA all compared notes and worked together more, but they don’t. It could be even more helpful to strengthen the CPSC. In 2021, Democratic Sen. Richard Blumenthal of Connecticut and Reps. Jan Schakowsky and Bobby Rush of Illinois introduced the Sunshine in Product Safety Act, which would aim to lift restrictions on the agency’s ability to warn people about dangerous products.
Under Section 6(b) of the Consumer Product Safety Act, the CPSC is barred from disclosing information about a manufacturer unless it has taken “reasonable steps” to make sure the information is accurate, fair, and reasonably related to the purposes of the CPSC. Practically speaking, it means the CPSC has to notify a manufacturer of a potential disclosure and give it time to comment, delaying the process of issuing a warning or recall and forcing it to negotiate with the manufacturer.
The Sunshine in Product Safety Act would repeal Section 6(b). “CPSC must be able to move swiftly to warn Americans when products like the Peloton Tread+ and the Fisher-Price Rock ’n Play Sleeper pose a danger to them and their families. Yet current regulatory constraints allow companies to call the shots on how and when to notify the public about their hazardous products, keeping important safety information from the public,” Blumenthal said in a statement at the time the legislation was announced.
Even when a recall is announced, it’s hard to get the word out. Consumers aren’t exactly checking the CPSC website every day for updates, and the media generally only picks recalls up when they’re really catastrophic or impact a lot of people.
Once a consumer product is sold on the market, it can be next-to-impossible to track. Say a parent buys a stroller when their child is born and later sells it at a garage sale, or, like in Ball’s case, someone fishes a dehumidifier out of a basement. Even if the manufacturer does alert people who bought the stroller or dehumidifier originally of the recall, that alert won’t make it to the person who has the item now. Fisher-Price has twice issued recalls on its Rock ’n Play sleepers after more infant deaths, even though the product has been pulled from the market. A 2021 USA Today investigation found products for sale on Facebook Marketplace that had been linked to the deaths of over 100 children.
There’s no one silver bullet solution to ensure all consumer products are safe. Some recalls are, to a certain extent, inevitable. But there are measures that would help.
To state the obvious here, it’s bad that unsafe products hit the market in the first place, and the more that can be done to slow that down, the better. “A lot of companies are not putting products through the proper testing. They might do their focus groups for marketing, but not testing,” Murray said.
Murray said she would like to see companies have to spend money on getting the word out on faulty products, especially given how much they spend on marketing those products to land the sale. “Companies spend bazillions of dollars targeting us to buy their stuff … we know what demographic they’re reaching,” she said.
And then there’s the issue of consequences — we don’t really have a good way to punish companies when they screw up. As the Washington Post noted in January, the CPSC has been more willing to hit companies with fines, including a $13 million penalty against retailer TJX Companies for knowingly selling recalled products and a $27 million penalty against Polaris Industries for not reporting defective off-road vehicles. These amounts, for these companies, are a drop in the bucket.
There are steps consumers can take to try to ensure the products in their homes are safe. If you’re looking to make a purchase, you can go to saferproducts.gov and drop in a brand name or product to see if there are any complaints. Every few months, it’s not a bad idea to check out the CPSC website to see whether any of the major items in your house have been recalled, especially those pertaining to your kids, or sign up for the CPSC’s email list to get recall alerts. We spend a lot of time looking up reviews to figure out which gadgets and restaurants are best; it’s worth taking a little time to make sure the major appliances in our homes aren’t about to burn them down.
Still, the whole recall system leaves a lot to be desired — and, again, companies often don’t want to say there’s a problem in the first place.
“People’s lives are often involved here, especially with drugs and devices, but definitely with cars and definitely with consumer products. I mean, I almost burned my house down with this dehumidifier that was years old that didn’t get recalled, and it should have been,” Ball said. “These things matter — how well products are recalled, how quickly — it matters a lot.”
We live in a world that’s constantly trying to sucker us and trick us, where we’re always surrounded by scams big and small. It can feel impossible to navigate. Every two weeks, join Emily Stewart to look at all the little ways our economic systems control and manipulate the average person. Welcome to The Big Squeeze.
"These amounts, for these companies, are a drop in the bucket." This is why we need penalties to be a % of a company's gross income. "You knowingly sold a product that was dangerous to the point that many people died from it? That's a 15% penalty. That eats all your profits for two years? Too bad, so sad." Companies who can absorb a $1 million fine will pay attention to something that actually hits their bottom line.
Newspapers across the country dropped the "Dilbert" comic strip over the weekend after the creator of the satirical cartoon went on a racist tirade, calling Black Americans a "hate group" and suggesting that White people should "get the hell away" from them.
About two decades ago, an important study found that roughly 1 in 4 Americans supported leaders who are uncompromising and take decisive action. These people said they would also prefer nonelected experts to make decisions. Our study replicates this finding nearly 20 years later but sheds light on a troubling reason for this preference.
At the Allegheny College Center for Political Participation, we, with our former student Candaisy Crawford, asked people about their willingness to support leaders who promised to protect them by any means necessary, even if that meant violating expected standards of behavior in a democracy, a set of principles often called “democratic norms.” We developed these questions based on existing research about the strategies that leaders with anti-democratic tendencies use to build public support.
In Venezuela, for instance, democratic decline happened gradually. Early on, Venezuela’s former president Hugo Chavez was known for using nationalist language and calling opponents epithets like “rancid oligarchs” and “squealing pigs.” Later, he blacklisted those who sought his removal from office through a democratically conducted referendum. Eventually, he went further, arresting and exiling his political opponents.
These types of tactics have also been used in other nations, such as Turkey and Hungary, by leaders who rose to power through democratic elections.
In our study, we asked about behaviors that foreshadow the early stages of democratic decline. For example, we asked citizens whether they thought that “the only way our country can solve its current problems is by supporting tough leaders who will crack down on those who undermine American values.” We also asked about explicit violations of democratic principles, like shutting down news organizations and “bending the rules to get things done.”
By design, some of these questions allow citizens to use their own interpretations of actions like “crackdowns” and “bending the rules.” These types of practices can take a number of different specific forms, as the cases of Venezuela, Turkey and Hungary illustrate. Our aim was to determine whether citizens were inclined toward leaders who seek power by promising retribution toward some groups and benefits for others, because this rhetorical strategy is often a precursor to explicit violations of democratic institutions.
Likewise, the phrasing of our questions is designed to allow respondents to rely on their own ideas about the meaning of “American values,” and “people like you.” Our interest was in what people would enable leaders to do to protect their idea of America and the Americans with whom they identify.
We found that people who want this type of protective but anti-democratic style of leadership were by far the most inclined to want leaders who would take uncompromising, decisive action. These people did not merely want their side to win a political competition for power. They were literally willing to say they would “bend the rules” to do it, a clear violation of the democratic ideal that everyone must follow the same rules.
For each item, we found that at least a third of the people we polled agreed or strongly agreed with these subtle or explicit violations of democratic norms.
While Republicans are certainly worse about this than Democrats, what negative polarization does is stoke the desire to see your opponents crushed and, quite honestly, dead. What surprised me here was not the high number of Republicans who believe this stuff but rather than surprisingly high number of Democrats who do too.
I *do* happen to support "tough leaders who will crack down on those who undermine American values.” I feel strongly that we need leaders who will deal with those who have tried to overthrow our country. Deal with them through the legal framework that our country has in place.