An awkward live TV moment for White House deputy chief of staff Stephen Miller is under heavy scrutiny on social media.

On Monday afternoon, the Donald Trump aide went on CNN’s News Central to discuss the President’s deployment of National Guard troops to U.S. cities.

[time-brightcove not-tgx=”true”]

CNN asked Miller whether the Trump Administration will abide by a district judge’s order blocking the Guard’s deployment in Oregon. “Well, the Administration filed an appeal this morning with the Ninth Circuit,” Miller began. “I would note the Administration won an identical case in the Ninth Circuit just a few months ago with respect to the federalizing of the California National Guard.”

Then, Miller said: “Under Title 10 of the U.S. Code, the President has plenary authority, has—” before making an abrupt stop. Miller blinked several times, with anchor Boris Sanchez calling out his name, though he still did not respond.

“Stephen, I apologize. It seems like we’re having a technical issue,” Sanchez said, before the show cut to a break, and resumed after with Sanchez saying some “wires got crossed.” The interview with Miller continued, but neither Sanchez nor Miller brought up “plenary authority” again. CNN did not immediately respond when asked by TIME for comment on what exactly the technical issue was.

The brief moment sparked controversy online over Miller’s use of the phrase “plenary authority,” which Cornell University’s Legal Information Institute defines as “power that is wide-ranging, broadly construed, and often limitless for all practical purposes.”

Some users speculated that there was no malfunction during the interview but rather that Miller abruptly caught himself and stopped talking. One TikTok user said, “Stephen Miller said the quiet part out loud accidentally and then appeared to glitch, all of a sudden when he recognized how much of his foot he just put into his mouth.” The user added: “He just came out and said Trump is authoritarian.” Another TikTok user posted a video with more than 100,000 likes: “Stephen Miller absolutely did not have a glitch on live TV. He had an ‘Oh sh-t’ moment where he said something that he f-cking was not supposed to say.”

On X, one user posted, “Stephen Miller didn’t glitch. He said ‘plenary authority’ and whoever was in his ear told him to STFU because he said too much and he froze like a deer in headlights.”

Upon return to air, Sanchez asked Miller about “the President’s legal authority as you see it under Title 10,” which refers to the part of the U.S. Code outlining laws that govern the Armed Forces, and “whether the Administration still plans to abide by that judge’s order, restricting any National Guard troops from being sent to Oregon.”

“Well, the Administration will abide by the ruling insofar as it affects the covered parties,” Miller said. “But there are also many other options the President has to deploy federal resources and assets under the U.S. military to Portland. But I was making the point that under federal law, Section— Title 10 of the U.S. Code, the President has the authority anytime he believes federal resources are insufficient to federalize the National Guard to carry out a mission necessary for public safety.”

A section of the law Miller mentioned states that the President may call into federal service members and a state’s National Guard under three conditions: if there is an invasion by a foreign nation, if there is a rebellion or a danger of one staged against the U.S. government, or if the President is “unable with the regular forces” to execute U.S. laws.

Trump has continued to test his executive authority in the deployment of troops within the U.S. and has been met with criticisms from local officials as well as pushback from court rulings, to which he has floated the possibility of using the 19th-century Insurrection Act to bypass.

Jared Decker was on his way to lunch when he received the phone call that kicked off a monthslong real estate nightmare.

It was the fall of 2023, and the caller was an aggrieved local businessman in Tampa, Florida, who had just discovered thousands of dollars missing from his bank account. The property management firm Decker works for, the caller said, had charged him nearly $10,000 for rent over several months — despite the fact that he had never lived at any of its more than 600 rental homes.

Decker and the owners of his company soon discovered the source of the mix-up. A fraudster had rented the home under a false persona, stealing the businessman’s bank account information and cribbing other identifying info, including his name, from a different person in Pennsylvania. The scammer had begun living at the property before Decker’s company, Coastal Pioneer Realty, took over managing it, so he hadn’t gone through their usual screening process, Decker tells me. The ensuing ordeal, including an eviction and court proceedings to sort out the missing money, ended up costing thousands of dollars in legal fees and missed rent payments. The property’s owner, a so-called “mom and pop landlord” who owns a handful of rentals, was on the hook for most of it.

Landlords and property managers nationwide are sounding the alarm over a surge in leasing application fraud, as scammers fabricate documents or steal entire identities to secure rentals. An executive at Greystar, the country’s largest multifamily property management company, tells me that about half of applicants in some pockets of Atlanta are rejected due to “provable” fraud. During an earnings call in 2023, the CEO of Camden Property Trust, another large apartment company, recounted his own brush with identity theft: A scammer, he said, tried to use his identity to lease an apartment owned by a competitor in Charlotte. The prevalence of these tactics is hard to track, especially since most of the hard data on it comes from business groups, leasing-software firms, or apartment operators that have clear incentives to make a fuss about fraud. But chat with any property manager and you’ll probably get some version of this rallying cry: Leasing fraud is a greater threat than ever.

Boo-hoo, you might think. Poor landlords. But dismissing leasing fraud as a champagne problem ignores the consequences for regular apartment hunters. As landlords beef up their crook-catching methods, they may also increase their reliance on imperfect tenant-screening software. Honest renters could face a tougher, more opaque application process as a result — and maybe even rent hikes. Even after they move in, tenants could wind up living down the hall from someone who isn’t who they claim to be. In other words, the panic over scammers is about to make renting more of a headache for everyone.

These tactics generally fall into two main buckets. There’s “first-party fraud,” in which prospective tenants apply using their real identity but spin up fake pay stubs, bank statements, or credit reports to score a unit for which they may not have otherwise qualified. Then there’s the more nefarious “third-party fraud,” which involves either stealing an identity wholesale or cobbling together a fake persona based on stolen information. Once a scammer has the keys, they may simply take up residence like any other renter, pocket some money by illegally subletting the unit to an unwitting tenant, or use the space as a “haven” for drug or sex trafficking, a report from the landlord-software company RealPage warns property managers.

We definitely have seen that leasing fraud attempts have grown in both volume and sophistication in recent years.Jamie Teabo, senior managing director at Greystar

Conditions have never been more ripe for a sophisticated scammer to flourish. The pandemic forced the industry to move leasing processes online, enabling fraudsters to hide behind screens. The advent of AI has also made it easier than ever to churn out fake documents that evade detection. This is especially perilous for small-time landlords, who own more than a third of rental units around the country and aren’t exactly known for their tech-savvy. Tens of thousands of newly built apartment units, many offering at least a month or two of free rent, have hit the market in recent years, leaving scammers with no shortage of targets. A survey of about 70 apartment operators conducted at the end of 2023 by the National Multifamily Housing Council, a trade group, found that 70% had seen an increase in fraudulent applications and payments over the preceding 12 months. A separate study last year by RealPage reported a similar rise — of roughly 400 property managers surveyed across five large metros, 75% reported an increase in fraudulent behavior. The NMHC survey also found that nearly 24% of evictions over the prior three years were tied to fraudulent applications and related failure to pay rent. The eviction process can take months, costing landlords both time and money.

“Five to seven years ago, fraud was, I don’t want to say nonexistent, but it was less of a factor,” says Chase Harrington, the president and COO of the property-management software company Entrata. Back then, he says, leasing offices were mostly on the lookout for applicants fudging their income or credit history to secure a place. “What we’ve seen progress over time,” Harrington tells me, “is now truly bad actors, of like, ‘Are they who they say they are?'”

Leasing fraud appears to be widespread; property managers I spoke to struggled to identify factors, such as location, type of building, or local tenant laws, that make places more vulnerable to the problem. They frequently cited Atlanta as a hub for this kind of activity, despite its location in a relatively landlord-friendly state with a speedier eviction process and fewer tenant protections. Greystar, which manages more than 960,000 multifamily units in North America and around 44,000 units in the Atlanta metro, says it flags around half of applications in some pockets of the city — most notably the midtown, downtown, and Buckhead areas —as fraudulent. Greystar said it identified 14% to 18.5% of applications as fraudulent in metros like Durham-Chapel Hill, North Carolina; Salt Lake City; Portland; Charleston, South Carolina; and Boston. These numbers don’t include applicants who drop out of the process when they encounter hurdles that are likely to ensnare scammers, like an ID check.

“We definitely have seen that leasing fraud attempts have grown in both volume and sophistication in recent years,” Jamie Teabo, a Greystar executive, tells me.

The purported rise in fraud has been a boon for companies that help rental owners identify and root out bad actors. One of these is Snappt, an identity- and income-verification software firm that raised a $100 million Series A funding round in 2022 and now conducts checks for landlords who own more than 2.2 million apartments across all 50 states. “Business has been kind of a rocket ship,” Kyle Nelson, the company’s vice president of corporate strategy, tells me.

Although Nelson says instances of fraud are trending higher than last year, Snappt’s tallies of fake application documents were significantly lower than the figures cited by Greystar. Snappt reviewed about 770,000 multifamily applications in the first half of this year and found an average fraud rate of 6.5%, with the highest volumes in Houston, Atlanta, and Los Angeles, all in the high single digits. Of the roughly 2.9 million documents the company analyzed, about 25,000 were deemed to be the result of so-called “template farms,” operations that mass-produce fake documents and enable applicants to “misrepresent their income, employment, or identity.” Another 21,540 were flagged as the work of “advanced fraud rings” that tamper with the underlying code of documents to evade manual detection.

It’d be one thing if this were merely a headache for landlords and management companies. But the people I spoke with in the rental industry argue that the rise of scammers is bad for existing tenants and potential renters, too. Apartments tied up in eviction proceedings are temporarily held off the market and out of reach of honest renters. Stricter screening methods mean regular applicants face more barriers to finding a place to live. Apartment operators already pass along a bevy of fees to renters — if fraud really does take a chunk out of their bottom line, odds are they won’t eat that bill, either. One obvious answer would be higher application fees to offset the cost of applicant screening.

Jay Parsons, a rental housing economist and consultant, describes fraud as the “big thing no one outside the industry wants to talk about.” With renters in the grips of an affordability crisis, there’s not much appetite for landlord hand-wringing. But brushing off the rise in fraud, he says, won’t help anyone.

“There are real challenges out there. I don’t want to downplay that,” Parsons tells me. “At the same time, we do see things like fraud that are actually exacerbating those challenges in the form of taking units offline that otherwise could be available.”

Screening has become just more and more opaque, and it’s often locking people out of the application process altogether, making it really difficult to access housingMarie Claire Tran-Leung, senior staff attorney at the National Housing Law Project

I couldn’t find any data that explicitly tied fraud to rent hikes — fraud is an already murky practice and just one cost that goes into a landlord’s decision-making. But when apartment operators miss out on months of rent or spend money to flush out fraudsters, “a lot of times they have to make up that money somewhere,” Caitlin Sugrue Walter, head of research and innovation at the NMHC, tells me. A 2022 survey by the Urban Institute backs this up: The nonprofit think tank found that landlords who missed rental income in the early days of the pandemic were more likely to raise rents and planned to jack up prices by a larger percentage, “indicating they are looking to recover some financial losses.”

Property managers may also tighten the screws on screening, the process of crawling through background checks, employment information, and credit histories to pick the tenants most likely to pay their rent on time and without issue. For renters, this screening is often a “black box,” Marie Claire Tran-Leung, a senior staff attorney at the National Housing Law Project, tells me — they may have no idea what their application fee is being used for, or which criteria might put them out of contention. A 2022 report from the Consumer Financial Protection Bureau identified error-riddled background checks and algorithmically derived “risk scores” that effectively reduced applications to a thumbs-up or thumbs-down.

“Screening has become just more and more opaque, and it’s often locking people out of the application process altogether, making it really difficult to access housing,” Tran-Leung says. “The application fees are of significant concern because people are having to pay a lot of money, in some cases, without a realistic chance of accessing that housing.”

A spokesperson for Snappt tells me in an email that its technology is built to “reduce bias, not reinforce it,” and that the company doesn’t decide who gets approved or denied for a place. The spokesperson also says that Snappt gives applicants flexibility in how they verify their information — through secure document uploads, for example, or linking directly to bank accounts or payroll providers.

“Traditional manual reviews and subjective decision-making can introduce human bias — sometimes unintentionally,” the spokesperson says. “By contrast, Snappt’s Applicant Trust Platform applies a consistent, evidence-based process to identify document tampering and verify income and identity data, without regard to race, gender, or background.”

Tran-Leung says the concern over leasing fraud is a distraction from the affordability crisis, a way for landlords to pass the buck on the rising cost of housing. Rents have increased by more than 25% since 2020, according to Zillow data, despite some recent easing due to a surge in new rental supply. Moody’s Analytics recently found that more than half of American renter households are considered “rent-burdened,” meaning they spend over 30% of their income on shelter.

That’s the tricky part of all this. Tenant advocates largely view landlords’ claims of rising fraud as overblown, a sideshow to the main issue of affordability. Housing operators say that the very real rise in fraud only exacerbates that problem. As technological advancements make distorting reality as simple as tooling around with a chatbot, I don’t think we’re close to seeing the last of these complaints.

Jared Decker, the property manager in Florida, certainly hasn’t. He says his office is dealing with another instance of fraud that summons memories of that quagmire back in 2023.

These kinds of instances are exceedingly rare, he tells me — usually the application process raises flags before a fraudster moves in. But as landlords and property managers guard against these nightmare scenarios, vying for an apartment has gotten more difficult for everyone else.

“It has become much more of a process, much longer, much more drawn out,” Decker tells me. “Many more requirements for the honest person because of all these fraudsters and scammers.”

James Rodriguez is a correspondent on Business Insider’s Discourse team.

• Jamie Dimon said JPMorgan’s $2 billion AI investment has already matched its cost in savings.
• “It’s the tip of the iceberg,” Dimon said of JPMorgan’s AI gains.
• His comments come as investors question if massive AI bets will really pay off.

JPMorgan CEO Jamie Dimon said the bank’s multibillion-dollar push into AI is already delivering results — and could just be the beginning.

Dimon said in an interview with Bloomberg TV on Tuesday that the bank spends about $2 billion a year on AI and is seeing about the same amount in direct benefits.

“We have shown that for $2 billion of expense, we have about $2 billion of benefit,” Dimon said. “We did this, we reduced headcount, we saved this time and money.”

“We know about $2 billion of actual cost saves,” he added. “It’s the tip of the iceberg.”

JPMorgan has been working with AI since 2012. It’s now embedded across nearly every part of the bank, from risk and fraud detection to marketing, customer service, and idea generation, Dimon said.

He also said JPMorgan’s in-house large language model, trained on internal data, is used by about 150,000 people each week.

“It’s quite productive,” he said. “Our managers and leaders have to do it.”

But the CEO didn’t sugarcoat the potential impact of AI on the workforce.

“People shouldn’t put their head in the sand. It is going to affect jobs,” Dimon said, adding that AI will enhance some aspects of work, but also eliminate some jobs.

“But you’re better off being way ahead of the curve and retraining people,” he said.

Dimon said the bank is focused on retraining and redeploying employees whose roles change due to automation. “We’ll have more jobs, but there’ll probably be less jobs in certain functions,” he added.

Dimon and JPMorgan did not respond to a request for comment from Business Insider.

Big AI bets are under scrutiny

Dimon’s comments come amid growing doubts over whether the massive corporate spending spree on AI is actually paying off.

Executives at Meta say they expect to spend $600 billion on AI infrastructure, including massive data centers, through 2028. OpenAI and Oracle have announced plans to put $500 billion into a data center project dubbed Stargate.

The staggering scale of those investments has fueled talk of an AI bubble and the potential for a pop that could bring the stock market crashing down from record highs.

A Goldman Sachs report published in June said that many firms pouring billions into AI have yet to see measurable gains, thanks to high infrastructure and compute costs.

“AI technology is exceptionally expensive, and to justify those costs, the technology must be able to solve complex problems, which it isn’t designed to do,” Jim Covello, the head of global equity research at Goldman Sachs, said in the report.

“The starting point for costs is also so high that even if costs decline, they would have to do so dramatically to make automating tasks with AI affordable,” he added. “In our experience, even basic summarization tasks often yield illegible and nonsensical results.”