I’m safer now in New York than nearly any other year I’ve lived here. Why don’t people believe it?
In 2005, the year I moved to New York City, the NYPD recorded 135,475 crimes in the seven “major felony” categories — homicides, assaults, rapes, and various forms of theft. The next year, that number dropped, and it kept dropping. By 2017, it was around 96,000, where it stayed until last year, when it snuck back over 100,000, mostly due to an increase in grand larcenies, auto and otherwise. The “non-major” felony data has followed a similar arc, dropping almost 30 percent since 2005. By the official crime numbers, the most dangerous year that I’ve lived in New York City was the first one — by a lot.
So why do people keep asking me if I’m safe here? Last week, I found myself assuring a kind older gentleman in town from Boston — where violent crime has been creeping up after pandemic lows — that he was almost certainly perfectly fine walking around the swanky restaurant-laden blocks near Times Square. Statistically, the Big Apple is safer than small-town America, and its per-capita murder rate was at the bottom of the list of big cities in early 2022. I’ve been intrigued and baffled hearing from people from my hometown (which leads upstate New York in crime statistics) who might have taken a day trip down to the city to see the Christmas decorations and eat some roasted chestnuts, but won’t travel here anymore because they think it’s a war zone.
Emotional stories speak louder than facts, perhaps especially in a city as storied as New York. Writing of the city’s crime narratives during a much more dangerous era, Joan Didion wrote of observers’ “preference for broad strokes, for the distortion and flattening of character and the reduction of events to narrative” — in other words, the nearly universal desire to make stories out of feelings, and then believe them. And when people ask me if “New York is safe,” they don’t want to know about numbers. They’re asking about feelings.
/cdn.vox-cdn.com/uploads/chorus_asset/file/24219582/1242290505.jpg)
Getty Images
How people perceive crime, and how politicians represent it to the electorate, has less to do with data and more to do with vibes. In October, while fact-checking the claims of rising violent crime that drove many midterm campaigns, the Pew Research Center’s John Gramlich noted that “the public often tends to believe that crime is up, even when the data shows it is down.” Data from the DOJ’s Bureau of Justice Statistics shows that there’s no increase in violent crime across the board in the US, and yet for most years in the last three decades, the majority of America adults thought there was more crime nationally than the year before, even though the opposite was true. Indeed, over three-quarters of those polled in October by Politico/Morning Consult said they thought violent crime was rising nationally and 88 percent said it was increasing or remaining the same in their own communities.
It’s not just ordinary citizens whose perceptions about crime in specific places can be markedly divorced from reality. In 1990, there were 2,262 murders in New York, and Mayor Eric Adams was a transit officer. But in May 2022, he claimed he’d “never witnessed crime at this level,” even though the murder total in 2021 was 488, a little less than a fifth of the 1990 level. (Per the NYPD’s own data, the rate of crime is more than 80 percent lower than in 1990.) Adams has alternated between saying he’s afraid to take the subway and taking credit for the city’s continued drops in violent crime. It’s hard to know what to believe, maybe even for Adams.
News coverage, which seizes on the stories that seem most emblematic of the problems the viewers and readers fear, has a profound effect on how people perceive criminal justice and their own safety. Furthermore, popular entertainment has seized upon crime as the most hardy and enduring way to draw in an audience, whether it’s the endless true crime documentary machine, the Lifetime movie engine, or the Law & Order empire. A steady diet of crime content magnifies our sense that crime happens at random all the time and that we’re the next target.
So even if the facts tell us that New York City, or maybe our hometown, is a safe place to live, that we’re highly unlikely to be the victim of a crime, and that within most of our lifetimes there’s been a marked change for the better, we still find ourselves living another story.
/cdn.vox-cdn.com/uploads/chorus_asset/file/24219595/1438801.jpg)
Graham Morrison/Getty Images
And that’s a dangerous place to be, something Didion detected so clearly in the city’s storytelling. “The imposition of a sentimental, or false, narrative on the disparate and often random experience that constitutes the life of a city or a country means, necessarily, that much of what happens in that city or country will be rendered merely illustrative, a series of set pieces, or performance opportunities,” she wrote in 1991, at a time when crime was truly, statistically skyrocketing. In her essay “Sentimental Journeys,” about the 1989 assault of a jogger in Central Park that led to the wrongful conviction of five teenagers from Harlem, she explored how that particular crime became a symbol of everything that was wrong with New York and, by extension, the country at large. Women had been assaulted and murdered in other contexts throughout the city, but the Central Park jogger case caught the imagination of the world primed to see it as illustrative of whatever they believed was wrong with humanity.
And so, while an actual woman was harmed, and five young men had decades of their lives taken from them by the state, politicians found a place to land their lofty rhetoric. Didion notes that “Governor [Mario] Cuomo could ‘declare war on crime’ by calling for five thousand additional police; Mayor Dinkins could ‘up the ante’ by calling for sixty-five hundred.” As if they’re crusaders in Gotham working with Batman to rid the town of crime, not public servants making decisions based on careful analysis. All these years later, history repeated: New York Gov. Kathy Hochul announced in September that cameras would be installed on 2,700 subway cars to “focus on getting that sense of security back” — a telling focus on feelings. At the same press conference, Adams said that “if New Yorkers don’t feel safe, we are failing.” Recently, subway conductors have begun announcing at nearly every stop that NYPD officers are on the platform “in case you are in need of assistance,” apparently as part of the police surge announced by Adams and Hochul.
Even if you buy that there’s a causal link between police presence and safety for subway riders, the new surge is cold comfort to those who actually ride the subways every day. An already beefed-up NYPD presence in and around subways has been obvious for well over a year; they were the only group of people who seemed to be able to walk around the transit system unmasked without fear of being fined. Cross through the busy Atlantic Avenue station in Brooklyn at 11 pm on a Thursday, and you’ll see clumps of three or four cops everywhere, chatting with one another as the traffic flows around them.
At the same time, there’s been a rise in killing in and around the subway (nine this year, rather than the pre-pandemic average of two per year; last Monday 3.5 million rode the subway in one day). When, in January, 44-year-old Michelle Go was pushed onto the tracks by a mentally ill man and killed, there were six cops in the station, and two nearby.
/cdn.vox-cdn.com/uploads/chorus_asset/file/24219619/1240915765.jpg)
Getty Images
And when there was an actual shooting in the subway last spring (thankfully resulting in no fatalities), the suspect remained at large for a day, with police unable to find him even though he left a credit card at the scene of the crime. Adams said one of the cameras in the station wasn’t functioning, as well as cameras in the stations before and after. The suspect was finally apprehended — by a civilian.
The point of all of this is that the narrative seems wrong, and that means the conclusion is off. The narrative goes like this: Crime is happening, and it feels like it’s happening more now than ever because people keep saying it is — even the mayor! Police deal with crime, and thus, we need more police, and they will stop the crime from happening. It is sentimental because it taps into feelings we have, and those feelings just seem like they’re true. But there are basic contradictions in the story. So continuing to tell it becomes a way of comforting ourselves, and also of increasing police budgets.
Yet the story doesn’t answer basic questions we ought to be asking: Why did that crime happen in the first place? What root problems does it illuminate, and how can those be solved? If the issue is, as Adams says, that there are many mentally ill people in the subways who commit crimes, are they New Yorkers who deserve a sense of protection and safety too? Does history show that increased police presence aids these people?
Or does “New Yorker” only refer to people like me?
The sentimental narrative simplifies the reality of the “disparate and often random experiences” of life and provides “performance opportunities.” The situation “offered a narrative for the city’s distress, a frame in which actual social and economic forces wrenching the city could be personalized and ultimately obscured,” wrote Didion. Cases like that of the Central Park jogger were a way for the city to deal with its general anxiety about the widening economic and social gap, something that had become stark during the 1980s, with the same solutions being suggested:
Here was a case that gave this middle class a way to transfer and express what had clearly become a growing and previously inadmissible rage with the city’s disorder, with the entire range of ills and uneasy guilts that came to mind in a city where entire families slept in the discarded boxes in which new Sub-Zero refrigerators were delivered, at twenty-six hundred per, to more affluent families …
If the city’s problems could be seen as deliberate disruptions of a naturally cohesive and harmonious community, a community in which, undisrupted, “contrasts” generated a perhaps dangerous but vital “energy,” then those problems were tractable, and could be addressed, like “crime,” by the call for “better leadership.”
That’s what I think about now when I listen to people tell me how dangerous New York is, or listen to the mayor give solutions — well, the same solution, over and over. The facts, and the problems, don’t match the “solutions”; they’re the conclusion to a story that lies on top of reality. It’s not that there’s nothing to worry about. It’s that, on the whole, we rush to the fix that assuages our fears, rather than hunting around for new ones. Or we change behaviors in order to protect ourselves from things that pose a very small threat, which gives us the emotional permission to ignore the much larger one we pose ourselves.
Amazon has quietly been developing AI software to screen job applicants.
Last week, Amazon extended buyout offers to hundreds of its recruiters as part of what is expected to be a months-long cycle of layoffs that has left corporate employees across the company angered and on edge. Now, Recode has viewed a confidential internal document that raises the question of whether a new artificial intelligence technology that the company began experimenting with last year will one day replace some of these employees.
According to an October 2021 internal paper labeled as “Amazon confidential,” the tech giant has been working for at least the last year to hand over some of its recruiters’ tasks to an AI technology that aims to predict which job applicants across certain corporate and warehouse jobs will be successful in a given role and fast-track them to an interview — without a human recruiter’s involvement. The technology works in part by finding similarities between the resumes of current, well-performing Amazon employees and those of job applicants applying for similar jobs.
The technology, known internally as Automated Applicant Evaluation, or AAE, was built by a group in Amazon’s HR division known as the Artificial Intelligence Recruitment team and was first tested last year. Amazon first built AI hiring technology in the mid-2010s but discontinued use of its system after it demonstrated a bias against women.
In an initial test, Amazon’s HR division believed that new machine learning models successfully guarded against biases based on race and gender, according to the internal document. Artificial intelligence has become more widely used in hiring across industries in recent years, but there remain questions about its role in introducing or amplifying biases that may occur in hiring processes.
An Amazon spokesperson did not provide comment before publication.
Amazon has for years invested heavily in trying to automate different types of work. In 2012, the company acquired a warehouse robotics company called Kiva, whose robots reduced the need for warehouse workers to walk miles on the job but simultaneously increased the pace and repetitiveness of their work.
Amazon has continued to research other ways to automate its warehouses and introduce new robots, in part because the company churns through so many front-line workers that it has at times feared running out of people to hire in some US regions. In its corporate wing, Amazon previously implemented an initiative called “hands off the wheel” that took inventory ordering and other responsibilities out of the hands of retail division employees and handed them over to technology.
Now, with the creation and expanded usage of the AAE technology, the roles of recruiters inside the second-largest private sector employer in the US could be altered permanently, potentially reducing the number of people Amazon needs to employ.
That is, when the company starts hiring again.
Amazon instituted a corporate hiring freeze earlier in the fall and, just last week, the New York Times reported that Amazon would lay off around 10,000 workers, or 3 percent of its corporate staff, in what would be the largest series of corporate job cuts in the company’s nearly three-decade history. Alongside layoffs in the company’s Alexa and Amazon gadgets divisions, the company sent buyout offers to large swaths of the company’s HR division, including all low- and mid-level recruiters in the US and India. If employees voluntarily walk away from their jobs, Amazon is offering three months of pay plus one week of salary for every six months of tenure at the company. These employees have to decide on the offer by November 29.
The division’s leaders said involuntary layoffs could still happen in the new year, depending in part on how many employees agree to leave the company voluntarily. Amazon CEO Andy Jassy also said that layoffs in the company’s core retail division would occur into 2023.
The AAE technology removes one key role that some recruiters serve at Amazon, which is evaluating job applicants and choosing which should move on to job interviews. The program uses the performance reviews of current employees, along with information about their resumes and any online job assessments they completed during their hiring process, to evaluate current job applicants for similar roles.
“[T]he model is achieving precision comparable to that of the manual process and is not evidencing adverse impact,” the 2021 internal paper read.
The technology was first tested on applicants for medical representative roles at Amazon, who work out of the company’s warehouse network. But since then, it has been used to select job applicants for roles ranging from software development engineers to technical program managers, opening up the possibility of future widespread use across the company.
Within the technology industry, there’s a realization that the Big Tech boom may be over. In many cases, pandemic-fueled business successes have reversed or plateaued. Now, tech titans like Amazon are looking to tighten their belts, seemingly in part by delivering on long-term bets that technology, and AI in particular, can do what humans do — and maybe more cheaply.
Issues at Tesla are piling up while its boss is distracted by his new company.
It’s the day before Thanksgiving, and Tesla is in trouble. Share prices for the electric carmaker are down more than 50 percent since the beginning of the year. And now, the company’s mercurial CEO, Elon Musk, is distracted by his shiny new $44 billion toy: Twitter.
Tesla has long benefited from a sterling reputation, both as a luxury EV maker and an investment opportunity. Yet issues at the company are racking up, and more customers and shareholders seem to be taking notice. In recent weeks, several posts critiquing Tesla’s build quality have caught attention on social media, and hundreds of thousands of Tesla cars have been hit with recalls. (Over-the-air updates will address problems with the vehicles’ tail lights and the front passenger airbag, two of the recent calls issued by the company.) So while Musk is busy putting out fires at Twitter — and running his other companies, SpaceX, the Boring Company, and Neuralink — Tesla’s reputation seems to be taking a hit.
Let’s start with Musk himself. In the wake of his Twitter acquisition, some experts and analysts are concerned that Musk’s new job might be undermining his responsibilities as CEO of Tesla and contributing to its cratering stock price. Keep in mind that Musk also financed much of his Twitter acquisition deal by selling off his own Tesla stock, and also reportedly authorized more than 50 engineers from Tesla to work at Twitter when he took over last month. At the same time, Tesla is currently facing a lawsuit that alleges that Musk’s compensation package in 2018 was inappropriately influenced by the Tesla board’s personal ties to Musk. (The lawsuit also calls Musk a “part-time CEO.”) Several lawsuits have been filed against the company related to its workplace, including lawsuits alleging sexual harassment, racism, and a “toxic” workplace culture, over the past year or so.
The cars themselves are having problems, too. Build quality has been a consistent criticism of Tesla, and a reliability study published by Consumer Reports this month found that the company continues to have issues with its body hardware and steering systems, among other issues. Repairing Teslas remains a major hurdle, as a Recode investigation illuminated this past summer. The challenge is so significant that GM claimed in a recent investor presentation that its dealers had apparently repaired more than 11,000 Tesla vehicles since last year. A TikTok video documenting the build quality of one Tesla, including a wobbling trunk lid, picked up more than 4 million views earlier this month.
Steven Elek, a data analyst for Consumer Reports, said in a statement, “Build quality continues to be an issue for Tesla. In our latest reliability survey, owners reported problems with body hardware, and paint and trim on the Model S, Model 3, and Model Y. Tailgates that don’t close properly, loose trim and molding, and faded paint are some of the specific defects we heard from Tesla owners.”
@car_connoisseur PLEASE. SOMEONE TELL ME WHY THIS IS SO BAD!?! This Plaid was over $120k!!! #tesla #teslatok #car #cars #carsoftiktok #funfacts
♬ Anti-Hero - Taylor Swift
Then, there are Tesla’s recalls, which total 19 since the beginning of this year (GM has issued 25, and Ford has issued 63, for reference, though they’ve also manufactured many more cars). In addition to the recalls last week, the company has also issued recalls to some vehicles over problems with their power steering in November and other cars that could possibly ignore stop signs in February.
Many of these recalls are issued without the cars experiencing wide-ranging safety problems, and they don’t require physical changes to the cars. Instead, they’re handled through over-the-air updates, which allow Tesla to make the necessary repairs through, essentially, internet downloads. These fixes are obviously easier to complete than taking a vehicle in for repair or replacing one entirely. Still, as Recode has previously explained, these kinds of recalls — because they’re so easy to address — could also create a cycle where regulators are constantly racing to catch up to dangerous software.
Tesla’s plan to become a self-driving car company doesn’t seem to be going so well, either. Even as other companies back away from their autonomous vehicle aspirations amid an economic downturn, Elon Musk is still touting Tesla’s so-called Full Self-Driving software — a beta version of the company’s system is supposed to become available to more Tesla owners by the end of this year.
And now Tesla is facing yet another lawsuit, filed in September by owners who say Musk has misled customers about how far away, and how functional, this technology is. The government has gotten involved, too: Reuters reported in October that the Department of Justice is investigating Tesla’s technology, and the Securities and Exchange Commission has also launched a probe. At the same time, an initiative called the Dawn Project is calling for a ban on Tesla’s Full Self-Driving software. To support that effort, the project has released videos and ads meant to highlight the dangers of Tesla’s technology, including a television campaign that shows a Tesla knocking down a child mannequin. The carmaker recently sent a cease and desist letter in response.
Oh, and Tesla’s competitors are starting to pose more of a threat. While the company is still the darling of the emerging EV industry, companies like Ford and GM are racing to get ahead of Tesla and scale up their manufacturing. Startups like Lucid and Rivian are also trying to beat Elon Musk’s cars in the luxury market.
These companies probably don’t mind that Musk is currently spending his time reinstating Donald Trump’s Twitter account and posting memes, since they could use the time to catch up. As Recode editor Adam Clark Estes wrote when news of Musk’s acquisition was first announced: “But if there ever was a time to get a head start, it’s now. Tesla’s boss is away. Time to play.”
This story was first published in the Recode newsletter. Sign up here so you don’t miss the next one!