To no small extent, Americans have grown accustomed to being watched.

For more than a decade, CCTV cameras have recorded a hefty portion of many people’s public movements, while ubiquitous smartphones have made us all perpetually vulnerable to unsolicited documentation (if not, social media notoriety).

Until recently, however, technical limitations constrained the reach of government surveillance (at least, in the physical realm). Police may have been swimming in CCTV footage. But using such videos to track any given person’s movements was an extremely time-intensive endeavor. To do so, an officer would need to sift through hours of footage captured on many different cameras.

Yet AI is now making it possible for the government to trace ordinary citizens’ movements with unprecedented ease and precision. Or at least, this is what the security firm Flock Safety has effectively promised — and the company’s critics have loudly warned. 

Flock provides police departments with networks of surveillance cameras — primarily, automatic license plate readers (ALPRs) that photograph each vehicle that passes them. 

The firm’s main selling point is also its chief political liability: Its ALPRs log not only license plate numbers, but also generate textual descriptions of all vehicles they detect and store this information in a nationally integrated, searchable database. This makes it technically feasible for police to monitor the travels of most any driver with a few keystrokes. 

Alas, some cops cannot be trusted with that power. In August, the Washington Post reported that at least 50 law enforcement officers had been charged or accused of using license-plate readers for illegitimate purposes, often to stalk their exes or romantic interests. Even more alarmingly, some police departments appear to have used ALPR data to surveil their own critics and protesters. 

These revelations fortified an already growing movement against license plate readers. In August alone, more than 90 cities and counties canceled contracts with Flock. And this backlash has been remarkably bipartisan. Deep blue cities like Los Angeles, Austin, Texas, and Cambridge, Massachusetts, have backed out of partnerships with Flock, while Republicans in Florida have barred local police from installing ALPRs on state highways, and the Texas GOP has paused funding for Flock cameras.

The anti-ALPR movement is responding to real problems with how that technology has been deployed. And yet, according to Flock and many police departments, these cameras are also helping to nab car thieves and locate missing persons.

All this raises the questions: Do ALPRs really have significant public safety benefits? And if so, is there a way to have our license plate surveillance and civil liberties too? 

To find out, I reviewed the latest academic research, dug through the worst horror stories in the news, and reached out to criminologists about the risks and rewards of ALPRs. Although the available evidence is limited, the technology appears to have modest — but genuine — crime-fighting benefits. And while the ALPR databases have been badly misused, this is in part because the tech has gone largely unregulated. With stringent safeguards, there may be a way to make it much harder for creeps and authoritarians to monitor law-abiding citizens’ comings and goings without forfeiting a useful law-enforcement tool.

The case for ALPRs

On a late summer day in Chamblee, Georgia, Leslie Barrera was pushing her 1-year-old son Mateo in a stroller when a car pulled up next to her. A man sprung out of the vehicle, pointed a gun at Barrera, and tried to grab the baby from his carriage. But he’d chosen his victim poorly: Barrera charged him, wrested away his gun, and tried to shoot the kidnapper, only for the firearm to jam. A woman (oddly, disguised as a man) then emerged from the car and snatched the child. The baby-nabbers proceeded to pile back into their vehicle and speed away.

With the aid of surveillance cameras, police quickly identified the car as a maroon Acura SUV. Unfortunately, the abductors had swapped out its valid license plate with a paper dealer tag. Decades earlier, that might have delayed, if not thwarted, the recovery effort. But this was 2020 — and Chamblee police had access to a Flock ALPR system. By searching through its database of vehicle photos, the cops discovered a shot of the same Acura with its registered metal license plate from days earlier. The plate led them to the suspect’s name and address — and thus, within just five hours of his abduction — to Mateo.

Flock has a lot of stories like this. The company boasts that their cameras “supported more than 1 million criminal investigations” and helped locate “more than 10,000 missing persons” last year. 

“Every single day, law enforcement across this country is using FLOCK technology to solve crime and find missing people faster,” Paris Lewbel, a spokesperson for Flock, told me. 

The limited evidence suggests that Flock’s claims about its cameras’ public safety benefits are overhyped.

But Flock’s colorful anecdotes and raw tallies tell us only so much. A Flock camera can be used in an investigation without making a decisive difference in its outcome. Without ALPR photos, Chamblee police may still have been able to tie the maroon Acura to Mateo’s kidnappers. 

To gauge Flock’s actual impact on public safety, we need (at a minimum) some quasi-experiments: When jurisdictions deploy advanced ALPR systems, do they see lower crime or higher clearance rates — both compared to their own past record and to simultaneous trends in Flock-less places?

There has been remarkably little research into these questions. But a recent study from the University of South Carolina’s Ian Adams and Scott Mourtgos sheds some light into them. 

Drawing on Flock’s installation records, the criminologists identified the month that 216 law enforcement agencies activated their first camera. They found that during the first year after Flock deployment, these areas saw an 11 percent decline in motor vehicle theft, relative to similar places that remained un-Flocked. Jurisdictions that put up the ALPRs also saw the success rate of their car theft investigations rise by 16 percent. 

These findings should be taken with a few grains of salt. For one thing, police departments will be more likely to invest in an ALPR system if they’ve got a serious stolen vehicle problem. And if car thefts in an area spike, they are liable to regress back toward historically typical levels, no matter what the local police department does. 

What’s more, agencies will often install ALPRs as part of broader anti-vehicle theft efforts, which may include deploying more officers to high-crime areas, among other things. Adams and Mourtgos’s data suggests that such simultaneous actions may be responsible for a portion of their results: Although Flock adopters solved stolen car cases at a higher rate after putting up the cameras, this rise tended to begin before the ALPRs were installed.

Nevertheless, the working paper indicates that Flock cameras do make some difference. “We’re not talking huge numbers here,” Mourtgos told me, “but it’s not nothing either.”

And there is reason to think that ALPRs have greater potential to deter and solve crimes than existing results suggest, as police departments may be deploying them in suboptimal ways. 

Consider a 2025 study of Atlantic City’s ALPR rollout. On paper, the municipality was an ideal candidate for the technology: As a barrier island, Atlantic City has only eight access points, so an ALPR network could easily capture every car entering and exiting the city. 

As one would hope, the research found that car theft declined after the cameras were erected, although the authors could not rule out other explanations. Perhaps more significantly, however, they discovered that the cameras produced vastly more leads than the department was capable of logging or investigating. Between November 2022 and November 2024, the city’s ALPRs issued 12,384 stolen vehicle alerts, virtually all of which went unnoticed or ignored by department personnel; over the studied period, the ALPR system prompted only 72 vehicle stops, of which just 52 concerned stolen cars. 

“In an ideal world, someone would be monitoring the alerts in real time and dispatching an officer, who would successfully make a traffic stop,” the paper’s co-author, Rowan University criminologist John Shjarback, told me. “But yeah, that kind of rarely happened.” 

Shjarback concludes that the department lacked enough patrol officers to respond to a significant portion of alerts, while its civilian surveillance center monitors appeared to simply miss a large proportion of serious hits. “They have a lot to do in that real-time crime center,” Shjarback said. “There’s a relatively high amount of crime and gun violence and whatnot in Atlantic City. One or two retired officers on duty just weren’t enough to do the live monitoring.” 

Similar operational flaws and resource constraints may have limited the efficacy of Flock systems in at least some of the areas analyzed by Adams and Mourtgos. Indeed, Mourtgos noted that many departments in his study lacked enough ALPRs to cover most significant thoroughfares. “Even in jurisdictions that are at the higher end on cameras, you’re still covering an extremely small minority of these roadway structures,” Mourtgos said. “So, that’s probably another reason why we see only modest effects.”

Taken together, the limited evidence suggests that Flock’s claims about its cameras’ public safety benefits are overhyped; its ALPR networks do not appear to be the “single greatest change” in policing tech “since DNA,” as the company touts on its website. But, as common sense might suggest, automatically logging all license plates on major highways and intersections seems to make it at least a bit harder to get away with stealing cars (and perhaps, babies).

Mass surveillance has its downsides

Of course, the fact that ALPRs offer some benefits to public safety does not mean that they’re necessarily worthwhile. Sewing a GPS chip into every American’s forearm would surely make crimes easier to solve and less tempting to commit. But most would regard such trackers as dystopian infringements on privacy and invitations to authoritarian control. 

And Flock’s ALPRS are much the same, in the view of their staunchest critics.

The “DeFlock” movement’s fundamental argument is that the cameras give police departments more information than they can be trusted with. One danger is that wayward cops can exploit the data to stalk their exes. The more profound threat, however, is that police agencies will use ALPR systems to deter or punish lawful dissent. 

And that hazard is not entirely hypothetical. After a writer named Canyen Ashworth published an op-ed criticizing local police in Lenexa, Kansas, that town’s police began tracking his movements with its ALPRs (apparently, because Ashworth’s column led the police chief to conclude, incorrectly, that he must have been the dastardly perp behind a recent spree of…unpermitted poster-hanging incidents).

More broadly, a 2025 investigation from the Electronic Frontier Foundation (EFF) found that more than 50 federal, state, and local agencies had run hundreds of searches related to protests or activist groups through Flock’s ALPR system. It’s not clear how many of these queries were motivated by an illiberal desire to monitor dissenters. In theory, a police officer could search for, say, “red sedan at No Kings protest” because a car of that description tried to run over protesters or committed some other criminal offense in the demonstration’s vicinity. And some of the searches tallied by EFF appear to fall into that category. But many others had no discernible purpose beyond the surveillance of activists.

We need to watch the watchmen

All this is unquestionably alarming. And if widespread misuse of ALPR networks were inevitable, then their costs to civil liberties would likely outweigh their modest (and as yet, not wholly certain) benefits for public safety.

But this probably isn’t the case. Rather, governments can likely limit ALPR’s downsides — while preserving their utility to law enforcement — by more stringently regulating the technology.

“In something like half of the states, there is no regulation. Virtually none, in our view, have adequate regulation.”

Barry Friedman, founding director of NYU School of Law’s Policing Project

Until recently, legislatures had placed scant guardrails on ALPRs. As the Congressional Research Service noted in 2024, there is “no specific federal legislative framework” governing law enforcement’s use of such cameras. And most states have no comprehensive ALPR regulations. Thus, if we’re failing to prevent police departments from misusing these cameras, it may be because most state and local governments haven’t actually tried. 

“In something like half of the states, there is no regulation,” Barry Friedman, the founding director of NYU School of Law’s Policing Project, told me. “Virtually none, in our view, have adequate regulation. The police decided to do this on their own. Them and their vendors made the decision and didn’t disclose what was happening.”

Fortunately, there is a long list of ways for policymakers to constrain abuse of ALPR data. To name just a few, governments could:

• Require police departments to delete logged location records after three weeks (or less), unless they are connected to a specific investigation (thereby leaving would-be stalkers and tyrants with less surveillance data to sift through). 
• Permit ALPR searches for only an enumerated set of law enforcement purposes, such as tracking stolen vehicles or suspected violent offenders. 
• Create an audit trail for every query documenting who searched for what, when, and under what justification.
• Mandate the use of software that automatically flags suspicious queries — such as someone searching for the same license plate over and over, without attaching an actual case number — and that temporarily locks out users who appear to be misusing the database. (Flock currently offers software with similar features, but most police departments with ALPRs are not legally required to adopt such systems.)
• Expressly forbid the use of the ALPR system for tracking people on the basis of protected speech.
• Punish intentional, unauthorized searches of the database with jail time.

Virginia’s recently enacted ALPR law includes many of these provisions, while several other states have embraced a subset of these policies.

Some may still want to get de-Flocked

To be sure, for some, even heavily regulated ALPR systems will entail intolerable costs to privacy and unacceptable risks of authoritarian misuse. The American Civil Liberties Union (ACLU) supports efforts to regulate license plate readers, but only as a second-best alternative to banning the technology outright. 

“No guardrails can completely address the chilling effect of mass surveillance,” Chad Marlow, senior policy counsel at the ACLU, said. “When people know that they are being surveilled all the time, they may not intellectually dig into the, ‘Well, what exactly is happening with this data? What’s it being collected for? How long is it being stored?’ The average person isn’t going to go through all those steps.” 

Rather, Marlow argues, many will simply see cameras and fear that they’re being watched — and fearful of those doing the watching. 

“Someone who is an alcoholic and wants to go to an AA meeting — but doesn’t want the public or any of their insurance companies or the government knowing they’re an alcoholic — and knows there’s a camera outside of the church that hosts the meeting may just say, ‘You know what? It’s really important for me to keep this struggle private. And so I can’t go to that AA meeting,’” he said. 

These are legitimate concerns, although there is little hard evidence that license-plate readers have a chilling effect on protests or personal behavior. And given how much sensitive information people publicly share on social media — or willingly generate, through their internet activity — there’s some reason to doubt that many will refrain from lawful activity, merely because they spotted an ALPR by the side of the road. In fact, license plate readers have actually been in wide use for more than a decade and, until recently, few seemed to register their existence, much less, reorganize their lives around avoiding the cameras’ gaze. 

Still, reasonable people can disagree about how to weigh ALPRs’ trade-offs. My point here is merely that governments may be able to secure modest but meaningful public safety benefits from the cameras — while greatly reducing their misuse — by giving regulation a chance. Using the ALPRs of 2026 probably doesn’t require accepting the authoritarian nightmares of 1984.

This story was originally published in The Highlight. To get access to member-exclusive stories like this every month, become a Vox Member today.

Todd Waller wasn’t expecting to spend the latter half of his 30s living with roommates. But in late 2022, following a breakup with his girlfriend, he needed to find a new place to live, quickly. 

A social worker based in the Atlanta metro, Waller started googling “efficiency apartments” and stumbled onto an ad for PadSplit, a company founded in 2017 that lets property owners rent out individual furnished bedrooms in single-family homes by the week. “It said I could get approved right away and move in tomorrow,” he recalled. There was no security deposit required.

All of a sudden, Waller found himself living with six other people and able to save a lot of money. His rent was just $180 per week, or less than $800 per month. Though he earned a decent salary, he never really had anything left over after all his bills were paid, and the new financial breathing room made the tradeoff of housemates feel worth it.

After a while, Waller decided he would be willing to upgrade to a nicer PadSplit unit, one with more amenities but still with roommates. Now he pays $1,250 a month to live on one side of a 2,000-square-foot home, with access to his own private bathroom, a private entrance, and a renovated kitchen. Three other people live on the other side of the house, and he rarely interacts with them.

“My house is awesome; I pay $500 less a month than I would if I were renting [a place of my own] without roommates, and I just was able to buy a new car,” he told me. He knows one day he’ll move somewhere else and perhaps even buy his own home, but he likes that he feels he can make that decision comfortably at his own pace.

With rents outpacing wages in most of the country, legislators are increasingly looking at co-living — sometimes called house-sharing or adult dorm rooms — as an overlooked piece of America’s housing stock. Zillow counts roughly 21 million “empty nest households” — or homes where older residents living with no children have at least two extra bedrooms. Some lawmakers want to make it easier for those homeowners to rent them out, which could put workers closer to their jobs and help seniors stay in houses they can no longer afford alone. Others are eyeing a different kind of vacancy: Post-pandemic remote work has left about a fifth of American office space empty, and there’s growing interest in turning some of it into apartments. 

Many advocates for people experiencing homelessness also argue that co-living is a needed solution, something that could help people going through emergencies avoid the scary slide into shelter living, especially models where you can move in almost immediately and without a deposit. PadSplit surveyed its renters in December and found a population that mostly works (83 percent were employed) and mostly hadn’t been to college. The median income was $32,500.

That co-living needs advocates at all is a legacy of the last century. Renting a single room in a boarding house or residential hotel was once an ordinary way to live in an American city, but in the decades after World War II those rooms became shorthand for poverty and vice, and cities responded by banning new ones and paying landlords to get rid of the old. Half a century later, many people still see sharing space with strangers as a last resort, or feel embarrassed about having roommates past a certain age. And discriminatory zoning codes written in that era are still on the books in most communities, enforcing traditional nuclear family arrangements when it comes to single-family housing, by defining “family” as people related by blood, marriage, or adoption. Those rules can be wielded against home-sharing, and against specific renters neighbors want to keep out.

But these longstanding barriers are starting to finally give. Doubling up with people you aren’t related to is less unusual than it used to be. Between 2017 and 2022, the number of households sharing space with nonrelatives grew by more than 500,000. Lawmakers desperate for anything that adds more housing supply also seem less picky about where it comes from. Companies like PadSplit are putting private capital behind the model, and a wave of new laws and litigation — much of it led by conservative legal groups arguing on property-rights grounds — is making it harder for cities to tell homeowners who can and cannot sleep in their spare rooms. 

For the first time since cities banned cheap rooms, the people trying to bring them back have the momentum, and the next couple of years will go a long way toward deciding whether co-living stays a niche workaround for people like Waller or becomes something a lot more people are able to consider.

The return of SROs

At the turn of the 20th century, as American cities swelled, immigrants, day laborers, and factory workers crowded into downtowns looking for somewhere cheap and temporary to sleep. A small room could be rented by the day, the week, or however long someone needed it. Some “boarded” in private homes while others took beds in bottom-rung hotels known as flophouses. These arrangements varied widely, but became broadly known as single-room occupancies (SROs).

The turn against them came after World War II, as middle-class white families decamped for roomy new houses in the suburbs, and SROs grew more closely associated with shoddy housing for the poor and the deviant. New York and other cities started banning SRO construction in the 1950s. By the 1970s, SROs were a regular media villain, even as they still housed hundreds of thousands of people, including the mentally ill who had been turned out by the closing of state asylums. Cities started to offer landlords incentives to convert their SROs into luxury rentals. Between the mid-1970s and the 1990s, roughly a million of these affordable market-rate units disappeared from the market across the country.

The idea that perhaps cities made a mistake in banning these lower-cost housing options has been simmering in urbanist policy circles for the last 15 years, particularly as the homelessness crisis has gotten worse. The Furman Center at NYU argued in 2018 that reintroducing SROs could help address the housing shortage. The architecture critic Karrie Jacobs argued in 2021 that cities had blamed a kind of housing for problems that belonged to the people living in it. “The notion took hold that if you eliminated the housing, the marginal people would simply vanish,” she wrote in Curbed.

Still, lawmakers have dragged their feet on actually making it easier to rent these kinds of units again, and until recently the people pushing for it had little to show beyond op-eds and white papers. What’s different now is that the fight has moved out of the world of policy papers and into places where it can produce actual results, including a lawsuit that could reach the Supreme Court.

Suing over who counts as family

In early 2022, a home-sharing startup called HomeRoom was renting out two houses in Shawnee, Kansas, a suburb of Kansas City. The company rented rooms one at a time to unrelated adults who wanted cheaper housing, and at a local city council meeting that year, Shawnee’s community development director asked whether the city should support such arrangements in single-family neighborhoods at all. Two months later, the council passed an ordinance banning “co-living groups,” which it defined as four or more unrelated adults sharing a home. 

Under the ordinance, if even one adult in a house wasn’t related to the others by blood, marriage, or adoption, the whole household counted as unrelated.

While the new law forced HomeRoom to evict its tenants, it also put Val French, a Shawnee homeowner, in a bad situation. She lived with her husband, their two adult sons, and one son’s girlfriend. Because the girlfriend was related to no one, the house was now deemed in violation of the city’s occupancy rules. Her son and his girlfriend had to move out, and French couldn’t rent the empty room to anyone else.

French and HomeRoom sued in federal court in May 2023, with assistance from the libertarian Pacific Legal Foundation. They argued that their right to choose who they live with is protected under the 14th Amendment, but they ran into legal precedent from a 1974 Supreme Court case, Village of Belle Terre v. Boraas, which upheld a Long Island village’s nearly identical rule against six college students sharing a house.

The decision came down in the same years cities were clearing out their SROs, and it has given local governments cover to decide who counts as a family ever since.

A federal judge dismissed the Shawnee suit in September 2023, citing Belle Terre, and this past April the US Court of Appeals for the 10th Circuit upheld the dismissal. A few months later, the Pacific Legal Foundation (PLF) petitioned the US Supreme Court to hear the case and reconsider Belle Terre once and for all. The precedent has already frayed in places — state courts in California, Michigan, New Jersey, and New York have struck down local “unrelated occupants” limits under their own laws — but it still governs everywhere else. 

There is “a fundamental liberty right to choose household living companions — whether as an articulation of the right of intimate association or independently — that is deeply rooted in American history and tradition,” PLF’s lead attorney on the case, David Deerson, told me. 

The Supreme Court hears only about 1 percent of the cases it’s asked to take, and even if it took this one, the plaintiffs would be asking a conservative majority to recognize a new constitutional right, which that majority has spent the past few years insisting courts should almost never do. 

Deerson’s argument is aimed at that skepticism, though. When the Supreme Court overturned Roe v. Wade in 2022, it held that an unwritten right counts only if it is “deeply rooted in history and tradition.” Deerson and his allies are arguing that living with people you aren’t related to qualifies as such because, for most of American history, boarding houses and lodgers were typical staples of family life. Belle Terre is also a short opinion written by William O. Douglas, a liberal justice whose legacy has little connection to today’s conservative justices. 

The legislative front

A Supreme Court reversal would be the cleanest fix for co-living, but advocates aren’t waiting on it. State legislatures have already shown they’ll act: Iowa barred cities from enforcing relationship-based occupancy limits in 2017, Oregon and Washington followed in 2021, and in 2024, Colorado lawmakers passed a law prohibiting cities from restricting how many unrelated roommates can live under one roof absent a clear health or safety reason. 

The Pacific Legal Foundation worked on that last bill too, and has since turned it into model legislation for other states. They colloquially call it the “Golden Girls bill” since occupancy caps like these would have prevented the four unrelated women of the hit 1980s sitcom from sharing a house. 

This year Pennsylvania lawmakers introduced their own version, with supporters citing not private property rights but the potential to ease the state’s affordable housing shortage. As legislators considered the proposal, Pacific Legal Foundation published research showing that nine of the 10 most populous cities in Pennsylvania capped the number of unrelated adults who could live together. The state House passed the Golden Girls bill this summer, though the Pennsylvania Senate version has not yet received a vote. 

The objection lawmakers hear most is the same one that animated Belle Terre — that lifting occupancy caps could turn quiet, orderly, neighborhood homes into de facto frat houses. That college partying concern killed Connecticut’s proposed Golden Girls bill this year. While the bill cleared the state Senate, it never got a vote in the House, amid reported concerns from the speaker about college students moving into wealthy neighborhoods. “What if a neighbor is unhappy about four unrelated people who are 19 years old or 22 years old living in the house?” state House Speaker Matt Ritter asked in an interview with the Connecticut Mirror. 

Pennsylvania’s sponsors tried to head off those fears by amending the bill to exclude protections where someone is enrolled at a college or university. In other words, the law would have covered Blanche, Dorothy, Rose, and Sophia from Golden Girls unless one of them signed up for a class at their community college.

Nevertheless, advocates are expanding their push. Jim Manley, state policy director for the Pacific Legal Foundation, told me he expects Texas, Florida, Pennsylvania, North Carolina, Rhode Island and Utah to take up new versions of the legislation in 2027.

And occupancy caps aren’t the only co-living target. Sam Hooper, a lawyer who has been lobbying state legislatures to re-legalize SROs, and his former colleagues at the Institute for Justice are now focused on the ROOM Act, model legislation that would bar cities from restricting SROs and other co-living arrangements, including rules that make it hard to convert offices into apartments. A version was first introduced in Rhode Island earlier this year. 

For Meagan Forbes, Institute for Justice’s director of legislation, the goal is for co-living to be treated like any other form of housing. “Other local laws can regulate things like noise and fire safety, but current regulations just go way too overboard in preventing this kind of housing,” she told me. 

Hooper, who drafted the ROOM Act, says he wanted to “take a massive sledgehammer” to existing restrictions. The resistance he runs into isn’t only the frat-house worry, but older cultural fears that SROs would recreate the poor tenement conditions of the early 20th century. “I think there’s this real problem of letting the perfect be the enemy of the good in housing policy,” he told me. “People accuse us of wanting people to live in undignified housing, but they just struggle to understand that it can be a perfectly fine option for many people at different points in their life.” Hooper credits his own London co-living experience in his 20s with introducing him to some of his best friends and his wife. 

Even so, he cited a broad political coalition backing him up, from property rights champions in conservative states to Democrats and progressive YIMBYs drawn to the housing justice and homelessness angles. There are business leaders interested in workforce housing, he said, and he noted that the AARP testified for the ROOM Act in Rhode Island. “I’m hopeful we’ll be able to keep this bipartisan coalition that hasn’t succumbed to the culture wars,” Hooper added.

Not every co-living proposal is about removing rules. DC’s Senior Co-Living Program Establishment Act, also nicknamed the Golden Girls bill, would pay senior homeowners $300 a month to rent a room to another senior, plus up to $2,500 for home modifications. 

The 21st Century ROAD to Housing Act, the federal housing package that became law in July, doesn’t touch occupancy rules. But it does fund a pilot program for converting vacant commercial buildings into housing, which could make adult dorm office conversions easier.

The demand question

While momentum for co-living is real, demand is also complicated. Americans are doubling up more than they were before the pandemic — about 8.2 million households in 2024, up 4.3 percent since 2019, according to analysis shared by Zillow. Yet the share made up of unrelated adults living together actually fell slightly, by roughly 50,000 households, while homes shared with couples or families grew by nearly 400,000. In other words, when Americans need to save on rent, they still mostly turn to people they already know.

PadSplit is betting that changes. Atticus LeBlanc recently announced his company is expanding into New York, Chicago, and San Francisco, three of the most expensive rental markets in the country.

“Does that mean things have gotten more desperate, that it’s forcing this cultural acceptance?” he asked me in an interview. “I think there’s certainly an argument to be made there. But you could just as easily make the argument that, holy cow, we are really, really wasteful with space.”