She introduced herself as Eve, but Ben knew right away that the voice on the other end of the line was a bot. Eve knew his name. She also knew how much money he’d owed a former landlord ($266). She didn’t seem to know that he’d settled with a collection agency five months prior. Eve said she was an AI agent from ProCollect and was calling to collect a debt. “Would you like to resolve it today by card or bank transfer?” she asked.

Ben had stepped outside on a balmy April afternoon in Portland, Oregon, to take the phone call. (He asked that WIRED use a pseudonym so he could speak freely about a financial issue.) As he stood in the sun, he wondered what he’d have to say to make Eve hand off a call to a human. “I figured it was just going to kick me over to a person when I asked about repayment structure or anything more technical,” he says. But Eve stayed on the line, so Ben did, too. He decided—why not?—to mess with the bot a little.

Ben says he asked the bot to engage in some role-play, in which he was “just a little guy” and his debt was like a giantess prone to trampling him. He wanted to see how weird Eve would get. The bot haltingly played along for a few minutes, he says, but then abruptly punted him to a call center employee. The human agent didn’t disclose whether they’d heard Ben’s bizarre conversation with the AI. They did, however, quickly clear up the confusion: “They looked me up in the system,” he recalls. “Found that the balance was zero.”

Ben’s experience is increasingly common. As inflation and stagnant salaries squeeze pocketbooks, debt delinquency in the United States is swelling. “We have, right now, the highest amount of collections in the courts that I've ever seen,” says debt settlement expert Michael Bovee.

“Hi, my name is Peter, and I’m a Claudeholic.”

It was August 2025 and Peter Steinberger was addressing a meetup in London called Claude Code Anonymous. Steinberger and some fellow addicts had arranged the event to network with people like themselves—techies swept up by coding tools such as Anthropic’s paradigm-busting Claude Code. “I dedicate pretty much all my waking time to this, yet it doesn’t feel enough,” he told the gathering in a cozy, brick-walled room.

A few months later, Anthropic released a new version of Claude Code, and the ranks of Claudeholics exploded. Called Opus 4.5, it could handle more complicated programming tasks, retain much more in its memory, run for many hours on end, and manage a team of AI subagents. Anthropic has what it describes as a “notoriously difficult” take-home exam for prospective engineering hires; in a head-to-head comparison of those people and its models, Anthropic claimed that Opus 4.5 “scored higher than any human candidate ever,” which “raises questions on how AI will change engineering as a profession.”

Countless coders spent the holidays in basements and dens, madly trying out this new toy that let them build software as if they’d unleashed a hundred clones. Or unlocked superpowers. “It feels like becoming Spider-Man,” one told me.

For the 39-year-old Steinberger, who split his time between homes in London and Vienna, even this was not enough. In November 2025, he launched a tool that’s now called OpenClaw, a simple way to conjure a personal AI agent that exploits the advances of Claude Code or other coding tools. Give it access to your data, your apps, and maybe even your credit card, and it scours your cloud and ventures onto the web to do your bidding. It can run autonomously in the background and overcome obstacles with the persistence of the Terminator.

Artificial Intelligence (AI) is now embedded in every aspect of every industry that matters. It can be a little intimidating. It can also be really exciting.

To help guide you through this transformation—which may or may not result in your redundancy—please complete this mandatory training webinar. You’ll be quizzed on key learnings throughout to test your compliance.* Don’t worry about being perfect. We know you’ll make mistakes. But that’s OK. You’re only human.

*Compliance does not guarantee your nonredundancy.

Awaiting Your Action
REQUIRED LEARNING DUE

Due Yesterday

NOTE: Training modules are designed using the latest multi-window technology and may include multiple overlays, types of content, and their own navigation. Good luck.

In the wake of attacks on CEOs, a nationwide protest movement targeting data centers, and increasing concerns about AI job replacement, federal intelligence agencies and domestic law enforcement are circulating reports with a new domestic target in mind: anti-technology extremists.

More than 1,000 pages of unpublished reports from the Department of Homeland Security, FBI, and fusion centers obtained by WIRED show a national shift taking place to surveil this new and worryingly broad category of people and activities deemed an emerging threat.

This new effort follows President Donald Trump's National Security Presidential Memo 7, which instructs the Department of Justice to target anyone holding “anti-American,” “anti-Christian,” and "anti-capitalism” beliefs. Earlier this month, Trump's counterterrorism czar, Sebastian Gorka, released a public counterterrorism strategy claiming that left-wing extremists are one of the three top counterterrorism priorities facing the United States.

Taken together, these Trump administration directives have commandeered the domestic surveillance apparatus to surveil and criminalize speech and assembly that challenges the ideology of the White House. A new focus on anti-technology extremism adds an unreported category to already public designations under a presidency that has heavily invested political and material capital in AI and data center proliferation.

As US voters look to the November midterms, the Trump administration is obsessed with looking back to past elections, seizing ballots cast years ago in several states in search, it claims, of fraud or other malfeasance. But experts believe the goal may be more varied.

The seizures began in January when FBI agents armed with a warrant raided an election facility in Fulton County, Georgia, and grabbed 600 boxes of ballots from 2020. This was followed in March by the Department of Justice obtaining ballot images from 2020 in Maricopa County, Arizona, and—citing claims about supposed fraud in 2020—demanding ballots from the 2024 election in Wayne County, Michigan.

These federal seizures have even trickled down to the local level. In March, a Republican sheriff in California obtained a warrant to seize about 650,000 ballots from a statewide redistricting election held in November. He announced, with no evident authority to do so, that his deputies would conduct a recount.

Election experts fear the trend could grow, creating widespread chaos after the midterms, if courts fail to scrutinize what appear to be politically motivated requests from groups intent on undermining election outcomes they don’t like.

“It’s really important for the public, for grand juries, for judges to not allow these actions … to become some kind of precedent,” says Gowri Ramachandran, director of elections and security at the Brennan Center for Justice. “This is not, and shouldn’t be, a rubber-stamp issue.”

It’s difficult to know for certain what parties seizing ballots aim to achieve. The DOJ could be fishing for evidence of fraud to legitimize President Trump’s claims that the 2020 election was stolen from him. Or it could be sending a message to voters and election officials that the federal government controls elections, despite the Constitution saying otherwise. The seizures may also be trial balloons to see how courts, election officials, and the public react. If the response is weak, it could embolden the administration to seize ballots after the midterms.

I algorithms exhibit racial bias in job candidate screening, and they discriminate more frequently against those applying for multiple jobs at different companies, according to Stanford-led researchers.

The boffins evaluated algorithmic hiring decisions across multiple employers that use the same hiring vendor. The resulting algorithmic monoculture, they say, is problematic.

The vendor in this instance was talent platform pymetrics, acquired by Harver in 2022. Harver did not immediately respond to a request for comment.

The researchers – Rishi Bommasani, Sarah H. Bana, Kathleen A. Creel, Dan Jurafsky, and Percy Liang – obtained a pymetrics dataset spanning the period from December 2018 through December 2022. It contained 4,197,168 job applications submitted by 3,372,132 applicants to 1,746 positions.

The dataset details hiring recommendations provided to 156 employers with a total annual revenue of $225 billion. It spans 11 industries, including finance, manufacturing, and warehousing.

When people applied for jobs at these companies, they were directed to pymetrics' machine learning platform to play assessment games. The platform's algorithm measures gameplay performance and recommends on average 58.2 percent of applicants per position. Employers decide who to interview, typically rejecting candidates who were not recommended by the hiring platform.

The researchers contend that the pymetrics algorithm is unfair.

"We find substantial evidence of racial disparities in AI-based candidate screening," the researchers said.

Carnival Corporation - the world's largest cruise operator - has confirmed a digital heist, a month after hacking crew ShinyHunters claimed to have stolen millions of customers' records.

The breach, Carnival confirmed, stemmed from an April 14 social engineering attack on an employee, though the company declined to comment on the scale or name ShinyHunters.

However, a company filing with the Maine attorney general's office puts the number of affected individuals at just under six million, down from the 8.7 million records previously listed by Have I Been Pwned.

Carnival previously acknowledged the phishing attack at the time, but it did not say whether any data had been accessed or stolen.

ShinyHunters claimed it lifted terabytes' worth of Carnival records and hinted at a breakdown in negotiations, likely related to the criminal outfit's extortion demands.

"The company failed to reach an agreement with us despite our incredible patience," ShinyHunters wrote on its data leak site, adding: "They don't care."

Are We Still at War?
Hahahaha.

At some point, everyone involved with the Iran war debacle, from political figures to Wall Street investors to the news media, will resolve to stop lunging at every Trumpian pronouncement of a “deal” to end the war. Even the best possible reading of the latest announcement cannot credibly be described as a deal at all, but rather, agreements to begin talks to reach a deal. The only specific outcome that would result is the opening of the Strait of Hormuz, which was open before the war began. It would not deal with the status of highly enriched uranium. It would not deal with Iran’s nuclear program. It would not deal with that nation’s missile production or capacity. It’s nothing, even in the best-case scenario, but a reset to the negotiations that were under way before the war began.

But we’re not in the best-case scenario, because that assumes that the non-deal deal is actually happening. Which it is not. Since this past weekend’s announcement by President Trump, we have seen Iran dispute really all of the terms of the deal and the U.S. admit that an actual agreement could take some time—which is what you say when nothing is agreed to. We have seen actual airstrikes on missile launch sites in southern Iran, which were described as “defensive” and which Iran has justifiably characterized as “flagrant” violations of the cease-fire agreement, while vowing to retaliate. (We haven’t seen that transpire yet, though there has been exchange of fire.)

"The core belief of the Trump administration in this and all things is that they can bully their opponent into submission."

When you are in a negotiation, and your negotiating partner violates the terms of the negotiations, you begin to get highly suspicious that any agreement you make will actually be adhered to. That is the position Iran finds itself in, and why we’re going to be in this endless loop of announcing an agreement to begin talks toward an agreement to end the war approximately indefinitely. Welcome to Groundhog Day.

Meanwhile, while everyone talks, correctly, about Iranian hard-liners at odds with their nation’s negotiating posture, Trump has had to attend to his own hard-liners who don’t want peace, here and in Israel, by making a rushed demand for every Arab country to formally recognize Israel, terming that a “complement” to a final deal. And Israel has been working assiduously to sabotage any agreement by bombing and even ordering evacuations in Lebanon, even as Iran is demanding an end to that conflict as one of the terms of the deal.