WIRED was able to download stories from publishers like The New York Times and The Atlantic using Poe’s Assistant bot. One expert calls it “prima facie copyright infringement,” which Quora disputes.

Poe, an AI chatbot platform owned by the question-and-answer site Quora and backed by a $75 million Andreessen Horowitz investment, is providing users with downloadable HTML files of articles published by paywalled journalistic outlets.

Prompting the service’s Assistant bot with the URL of this WIRED story about the AI-powered search service Perplexity plagiarizing one of our stories, for example, yields a detailed, 235-word summary and a 1-MB file containing an HTML capture of the entire article, which users can download from Poe’s servers directly from the chatbot.

WIRED was similarly able to retrieve articles from paywalled sites including The New York Times, Bloomberg Businessweek, The Atlantic, Forbes, Defector, and 404 Media in downloadable format simply by entering URLs into the Assistant bot’s interface. This appears to be just the latest example of the AI industry’s cavalier approach to intellectual property law, which is rapidly undermining existing business models in fields like journalism and music.

“This is a significant copyright issue,” James Grimmelmann, professor of digital and information law at Cornell University, wrote in an email. “Because they made a copy on their own server, that's prima facie copyright infringement.” (Quora disputes this, comparing Poe to a cloud storage service.)

When asked to summarize the content of a test website controlled by my colleague Dhruv Mehrotra, the bot did not return a summary but did return an HTML file. According to the website’s server logs, immediately after the Assistant bot was prompted to summarize the site, a server identifying itself as “Quora Bot” visited the site. It did not attempt to visit the site’s robots.txt page, suggesting that Poe and Quora ignore the Robots Exclusion Protocol, a widely accepted though not legally binding web standard.

A prominent media executive, whom WIRED granted anonymity to candidly discuss a legally sensitive matter his company is actively investigating, says that his publication also observed servers identifying themselves as Quora bots accessing its site immediately after giving Poe’s chatbot prompts about specific articles; these prompts, he says, yielded much or all of the text of these articles.

“Poe is a platform that lets users ask questions and have back-and-forth dialog with a variety of AI-powered bots provided by third parties,” Quora spokesperson Autumn Besselman wrote in an email. “We do not have or train our own AI models. Poe has a feature that enables a user to show the contents of a URL to a bot, but the bot will only see content that it is served by the domain. We would be happy to connect with your technical team to help them make sure your paywalled content isn’t served to people using Poe.”

“The file attachments on Poe are created at the direction of users and operate similarly to cloud storage services, ‘read it later’ services, and ‘web clipper’ products, which we believe are all consistent with copyright law,” Besselman wrote in response to an email asking follow-up questions. Andreessen Horowitz did not respond to a request for comment.

Quora cofounder Adam D’Angelo is a former Facebook CTO as well as an OpenAI board member. (D’Angelo was among the board members who voted to fire OpenAI CEO Sam Altman last year and was the only member who subsequently joined a new board formed after Altman’s return.) In March, nearly three months after Andreessen Horowitz announced that it had led a funding round for Quora, he sat for an interview with David George, a general partner at the venture firm who asked about the relationship between Quora and Poe.

“We’d love to have all of this as integrated as possible,” said D’Angelo. “If you think about the relationship between Facebook and Facebook Messenger, these are two products built by the same company, but they share a lot. I think that Poe and Quora might evolve to a similar kind of relationship. We’d love to get more of the human aspects of Quora into Poe. We’d also love to get the whole Quora dataset into the Poe bots and we’re also working—we’ve launched some of this already—to get some of the Poe AI to generate answers that are available on Quora.”

A user who goes to Poe’s website is prompted to start a chat with one of a number of chatbots. These range from bots offering direct access to various large language models and specialized bots like Mrstherapist (“Your very own therapist with relationship, trauma, stress etc.”) to the Assistant bot, the default option.

The Assistant bot is designated on the site as “by Poe” and as powered by AI firm Anthropic. This doesn’t mean that the bot is Anthropic’s; as the company points out, customers with access to its API can implement its models in their own products however they see fit. Claude, the Anthropic model on which the Assistant bot is apparently built, does not have access to the internet but can work with text provided to it—in this case, seemingly, by Quora crawlers that scrape websites in response to prompts. The HTML file the Assistant bot offers for download after being prompted with an article’s URL re-creates the article when opened in a browser, making it functionally equivalent to a PDF file.

Journalism companies have threatened and carried out legal action over claims that AI companies are infringing on their copyrights. The New York Times, for example, is suing OpenAI and Microsoft alleging infringement, and Forbes reportedly recently sent Perplexity a letter accusing it of “willful infringement.” (OpenAI, Microsoft, and Perplexity have denied wrongdoing.) Publishers whose articles WIRED was able to download were furious when apprised of the situation.

“As the law and The Times’ terms of service make clear,” wrote Charlie Stadtlander, a spokesperson for The New York Times, in an email, “scraping or reproducing The Times' content is prohibited without our prior written permission.”

“These stupid chatbots drive me fucking crazy,” wrote Defector co-owner and editor in chief Tom Ley in a text message. “Defector does not condone having its precious blogs stolen by a dumb chatbot backed by egghead-ass Andreessen Horowitz.”

Quora’s Chatbot Platform Poe Allows Users to Download Paywalled Articles on Demand

More than a dozen men threatened, assaulted, tortured, or kidnapped 11 victims in likely the worst-ever crypto-focused serial extortion case of its kind in the US.

Cryptocurrency has always made a ripe target for theft—and not just hacking, but the old-fashioned, up-close-and-personal kind, too. Given that it can be irreversibly transferred in seconds with little more than a password, it's perhaps no surprise that thieves have occasionally sought to steal crypto in home-invasion burglaries and even kidnappings. But rarely do those thieves leave a trail of violence in their wake as disturbing as that of one recent, ruthless, and particularly prolific gang of crypto extortionists.

The United States Justice Department earlier this week announced the conviction of Remy Ra St. Felix, a 24-year-old Florida man who led a group of men behind a violent crime spree designed to compel victims to hand over access to their cryptocurrency savings. That announcement and the criminal complaint laying out charges against St. Felix focused largely on a single theft of cryptocurrency from an elderly North Carolina couple, whose home St. Felix and one of his accomplices broke into before physically assaulting the two victims—both in their seventies—and forcing them to transfer more than $150,000 in Bitcoin and Ether to the thieves' crypto wallets.

In fact, that six-figure sum appears to have been the gang’s only confirmed haul from its physical crypto thefts—although the burglars and their associates made millions in total, mostly through more traditional crypto hacking as well as stealing other assets. A deeper look into court documents from the St. Felix case, however, reveals that the relatively small profit St. Felix’s gang made from its burglaries doesn’t capture the full scope of the harm they inflicted: In total, those court filings and DOJ officials describe how more than a dozen convicted and alleged members of the crypto-focused gang broke into the homes of 11 victims, carrying out a brutal spree of armed robberies, death threats, beatings, torture sessions, and even one kidnapping in a campaign that spanned four US states.

In court documents, prosecutors say the men—working in pairs or small teams—threatened to cut toes or genitalia off of one victim, kidnapped and discussed killing another, and planned to threaten another victim’s child as leverage. Prosecutors also describe disturbing torture tactics: how the men inserted sharp objects under one victim's fingernails and burned another with a hot iron, all in an effort to coerce their targets to hand over the devices and passwords necessary to transfer their crypto holdings.

“The victims in this case suffered a horrible, painful experience that no citizen should have to endure,” Sandra Hairston, a US attorney for the Middle District of North Carolina who prosecuted St. Felix’s case, wrote in the Justice Department's announcement of St. Felix's conviction. “The defendant and his coconspirators acted purely out of greed and callously terrorized those they targeted."

The serial extortion spree is almost certainly the worst of its kind ever to be prosecuted in the US, says Jameson Lopp, the cofounder and chief security officer of Casa, a cryptocurrency-focused physical security firm, who has tracked physical attacks designed to steal cryptocurrency going back as far as 2014. “As far as I'm aware, this is the first case where it was confirmed that the same group of people went around and basically carried out home invasions on a variety of different victims,” Lopp says.

Lopp notes, nonetheless, that this kind of crime spree is more than a one-off. He has learned of other similar attempts at physical theft of cryptocurrency in just the past month that have escaped public reporting—he says the victims in those cases asked him not to share details—and suggests that in-person crypto extortion may be on the rise as thieves realize the attraction of crypto as a highly valuable and instantly transportable target for theft. “Crypto, as this highly liquid bearer asset, completely changes the incentives of doing something like a home invasion," Lopp says, “or even kidnapping and extortion and ransom.”

**Three Failed, Brutal Crypto Extortions**

Despite the potential of stealing crypto through that kind of physical coercion, St. Felix’s gang had remarkably little success extorting significant sums of cryptocurrency from victims. But it wasn't for lack of trying: One document written by prosecutors, which outlines the basis for a plea agreement for one of the gang’s lookout drivers, describes how the group began to form in 2021 and went on to carry out a series of brutal attacks involving 13 total members of the gang over the next two years in at least seven actual or planned operations in Florida, Texas, North Carolina, and New York, all targeting victims they believed to have large stashes of crypto.

The group's string of crypto burglaries had its origins, in fact, in more traditional hacking-based crypto theft. According to prosecutors, one member of the group, Jarod Seemungal, and several alleged associates he'd met via the online game _Minecraft_, began in late 2020 to carry out so-called SIM swaps—in which hackers trick a phone company into transferring a victim's phone's service to their own device and then steal their two-factor authentication codes—to gain access to victims' online accounts and siphon out their crypto funds. Prosecutors say Seemungal's crew carried out multiple SIM swaps, prosecutors write, including one case in which they stole more than $3 million from a single victim.

By 2022, however, Seemungal's gang had begun to consider expanding their tactics to a far more violent form of crypto theft in order to hit targets who couldn't easily be robbed via hacking. He approached three members of what would eventually become the wider burglary group—St. Felix as well as two other young men in South Florida. St. Felix then went on to recruit most of the rest of the dozen-plus accomplices, all of whom coordinated their crimes on the messaging service Telegram.

In their first break-in, according to the prosecution's plea document, the group targeted the same victim from whom Seemungal had already stolen more than $3 million via SIM swapping, seeking to steal another $500,000 in crypto that she had managed to retain. At 11:30 pm on September 12, 2022, St. Felix and at least one other member of the group, wearing masks and armed with handguns and a rifle, broke into the woman's living room by shattering a sliding glass door. After struggling with the victim and another member of her household who suffered from Parkinson's disease, they put the woman on her knees, held a gun to her head, and demanded the password to an account on the Gemini crypto exchange.

She refused to give up her password, and was, according to the prosecutors' description, so demoralized by the earlier hacking theft of the majority of her funds that she told the men to simply shoot her. Instead, they stole her engagement ring, two iPhones, a laptop, the charger for the neurostimulator used by the other member of the household as a treatment for Parkinson's disease, and whatever cash they could find, then left.

For their next victim, the prosecutors describe how the group targeted a man who Seemungal knew to be a fellow SIM swapping hacker and who he believed had in fact robbed him of a significant sum of cryptocurrency in 2021. To prepare for that robbery in September of 2022, they began repeatedly sending their target pizza deliveries in the hope of conditioning him to come to his door without suspicion. When the moment of their planned theft came, however, their target wasn't home, so they instead lay in wait, then drew guns on their target when he arrived at the house.

Over the next hour, the group bound their victim's hands behind his back with bootlaces and demanded he hand over access to his crypto accounts. When the account he gave them access to had only a small sum of crypto, they put him in the backseat of their rented Cadillac, struck his face with their guns, drove away, and began extorting his friends and father for crypto payments. Eventually, about 120 miles from their victim's home, the men took their victim out of the car and told him to kneel. He instead escaped, as one of the men fired a gun from the moving car, though it's not clear if the shot was intended to hit the victim or merely scare him. One of the group—who has not yet been charged—would later say that St. Felix had suggested they kill their captive.

A few months later, prosecutors write, the group carried out their next attack against another victim they believed to be a wealthy crypto-focused hacker, this time in Texas. On a road trip from Florida to start surveilling their target, St. Felix had fled from law enforcement in Louisiana, flipped his car at more than 90 miles per hour, and broken his leg. The other members of the Florida crew had been arrested after that crash. So the break-in was carried out by a newly recruited team based in the Houston area.

Just a few days before Christmas of 2022, the Texas group broke into their target's home, bound his family members' hands with zip ties, and repeatedly hit him in the face demanding he give them access to his cryptocurrency. Prosecutors say they shoved knives and forks under his mother's fingernails and struck her in the face with a gun. They burned their target's arm with a hot iron to coerce him to hand over his crypto account details, and at one point attempted to punch him in the genitals.

The victim eventually told his torturers that he had buried a device storing his cryptocurrency in the backyard. (In fact, that hardware wallet, holding $1.4 million in crypto, was in a moving box in the home that the thieves never found.) When the thieves brought their victim to the backyard to locate the device, he climbed a fence and escaped. The burglars stole $150,000 in cash as well as some jewelry, then left.

**One Final Job**

In early 2023, after those relatively unsuccessful attempts at extortion, an associate of Seemungal's allegedly began feeding the group tips, hacking into potential targets' email to see the size of their crypto holdings, and sending those leads to the home invasion crew. One Telegram chat obtained by prosecutors shows a discussion of potential targets, including someone with $1.2 million in Texas and another person with $600,000 in Tennessee.

A screen capture of the group's Telegram chat as they discussed potential targets. A “lick” here is slang for a robbery target.

Courtesy of Dept. of Justice

They decided to go after a "new targ in north carolina” that they described as “some old dude”—a crypto investor in his seventies. In April, St. Felix and his crew drove to Durham, North Carolina, and knocked on the victim's door at 7:30 am, dressed as construction workers. When his wife came to the door, they knocked her to the floor. When the elderly man heard his wife scream and entered the room, they punched him in the face, bound the couple with zip ties, and dragged his wife into a bathroom.

During the rest of their time in the couple's home, they coerced the man to give them access to his crypto holdings by threatening to cut off his toes and genitalia, to sexually assault his wife, and to shoot him. The man eventually transferred about $156,000 worth of crypto from his Coinbase exchange account to his captors before Coinbase began to block his transactions. The thieves then destroyed their victims' computers and phones, left them bound with zip ties in the bathroom, and escaped.

That extortion incident would be the group's last, according to the prosecutors. On another occasion, they broke into a would-be victim's home only to discover it was an empty rental property. St. Felix and Seemungal were still planning more attacks in July of 2023 when St. Felix was arrested in the parking lot of a McDonald's in West Hempstead, New York, with an AK-style rifle and zip ties in his vehicle. According to the group's communications, they had discussed another target in Orlando, Florida, and yet another in Long Island, New York. For that final victim, the group had hacked the target's email, and St. Felix had suggested using one of the person's three children for leverage.

**Should Have Stuck to Hacking**

The criminal complaint against St. Felix suggests that identifying him as the ringleader of the crypto-focused home invasion crew wasn't particularly hard. Cell tower records placed his phone at the location of the Durham break-in. Bank records show his checking account purchases from Walmart of the construction worker outfit he would wear during the robbery. Google cloud storage records show screenshots of the reconnaissance he carried out in targeting his Durham victim.

The FBI's analysis of St. Felix's cryptocurrency transactions found that he used the instant exchanger FixedFloat to swap bitcoin and ethereum stolen from the North Carolina man for Monero, a cryptocurrency designed to be difficult to trace, and then back to ethereum. But because he carried out those swaps in a single web session, prosecutors were able to subpoena FixedFloat for its records and show that the sender and recipient were almost certainly the same person, despite his attempts at obfuscation. “FixedFloat basically connected the beginning and end points of the tunnel,” says Chris Janczewski, head of global investigations at cryptocurrency tracing firm TRM Labs, which published an analysis of the case. “It doesn't matter what you do in between if you can connect the inputs and outputs.”

Overall, the risk-reward trade-offs of carrying out brutal home invasions like the ones in St. Felix's case don't make nearly as much sense as remote hacking for crypto theft, Janczewski points out, although both tactics are highly illegal and potentially ruinous for victims. St. Felix and his coconspirators made a relatively small amount of money in the long run, carried out highly risky physical intrusions, and now face years in prison. St. Felix himself, for instance, faces a mandatory minimum of seven years and a maximum of a life sentence.

Still, crypto holders should nonetheless beware of the apparent rise in physical coercion as a means to steal their funds, says Casa's chief security officer, Jameson Lopp. He offers tips for protecting yourself against those attacks that include maintaining privacy—thieves won't go after a crypto stash they don't know about or can't physically locate—and creating technical hurdles to quickly sending large sums of crypto, even for yourself.

“If you can spend millions of dollars in a few seconds or a few minutes, then that's a pretty good indication you probably don't have enough layers of security around those assets," Lopp says.

Cryptocurrency's appeal, of course, has long included that it allows instant, frictionless transfers of vast amounts of wealth. Any crypto holder unlucky enough to be faced with a threat to their life or their loved ones, however, may quickly find themselves wishing for a form of money that didn't hold the same appeal to the very dangerous person putting a gun to their head.

Inside a Violent Gang's Ruthless Crypto-Stealing Home Invasion Spree

AWS hosted a server linked to the Bezos family- and Nvidia-backed search startup that appears to have been used to scrape the sites of major outlets, prompting an inquiry into potential rules violations.

Amazon’s cloud division has launched an investigation into Perplexity AI. At issue is whether the AI search startup is violating Amazon Web Services rules by scraping websites that attempted to prevent it from doing so, WIRED has learned.

An AWS spokesperson, who talked to WIRED on the condition that they not be named, confirmed the company’s investigation of Perplexity. WIRED had previously found that the startup—which has backing from the Jeff Bezos family fund and Nvidia, and was recently valued at $3 billion—appears to rely on content from scraped websites that had forbidden access through the Robots Exclusion Protocol, a common web standard. While the Robots Exclusion Protocol is not legally binding, terms of service generally are.

The Robots Exclusion Protocol is a decades-old web standard that involves placing a plaintext file (like wired.com/robots.txt) on a domain to indicate which pages should not be accessed by automated bots and crawlers. While companies that use scrapers can choose to ignore this protocol, most have traditionally respected it. The Amazon spokesperson told WIRED that AWS customers must adhere to the robots.txt standard while crawling websites.

“AWS’s terms of service prohibit customers from using our services for any illegal activity, and our customers are responsible for complying with our terms and all applicable laws,” the spokesperson said in a statement.

Scrutiny of Perplexity’s practices follows a June 11 report from Forbes that accused the startup of stealing at least one of its articles. WIRED investigations confirmed the practice and found further evidence of scraping abuse and plagiarism by systems linked to Perplexity’s AI-powered search chatbot. Engineers for Condé Nast, WIRED’s parent company, block Perplexity’s crawler across all its websites using a robots.txt file. But WIRED found the company had access to a server using an unpublished IP address—44.221.181.252—which visited Condé Nast properties at least hundreds of times in the past three months, apparently to scrape Condé Nast websites.

The machine associated with Perplexity appears to be engaged in widespread crawling of news websites that forbid bots from accessing their content. Spokespeople for The Guardian, Forbes, and The New York Times also say they detected the IP address on its servers multiple times.

WIRED traced the IP address to a virtual machine known as an Elastic Compute Cloud (EC2) instance hosted on AWS, which launched its investigation after we asked whether using AWS infrastructure to scrape websites that forbade it violated the company’s terms of service.

Last week, Perplexity CEO Aravind Srinivas responded to WIRED’s investigation first by saying the questions we posed to the company “reflect a deep and fundamental misunderstanding of how Perplexity and the Internet work.” Srinivas then told Fast Company that the secret IP address WIRED observed scraping Condé Nast websites and a test site we created was operated by a third-party company that performs web crawling and indexing services. He refused to name the company, citing a nondisclosure agreement. When asked if he would tell the third party to stop crawling WIRED, Srinivas replied, “It’s complicated.”

Sara Platnick, a Perplexity spokesperson, tells WIRED that the company responded to Amazon’s inquiries on Wednesday and characterized the investigation as standard procedure. Platnick says Perplexity made no changes to its operation in response to Amazon’s concerns.

“Our PerplexityBot—which runs on AWS—respects robots.txt, and we confirmed that Perplexity-controlled services are not crawling in any way that violates AWS Terms of Service,” Platnick says. She adds, however, that PerplexityBot will ignore robots.txt when a user enters a specific URL in their prompt—a use-case Platnick describes as “very infrequent.”

“When a user prompts with a specific URL, that doesn’t trigger crawling behavior,” Platnick says. “The agent acts on the user’s behalf to retrieve the URL. It works the same way as if the user went to a page themselves, copied the text of the article, and then pasted it into the system.”

This description of Perplexity’s functionality confirms WIRED’s findings that its chatbot is ignoring robots.txt in certain instances.

Digital Content Next is a trade association for the digital content industry whose members include The New York Times, The Washington Post, and Condé Nast. Last year, the organization shared draft principles for governing generative AI to prevent potential copyright violations. CEO Jason Kint tells WIRED that if the allegations against Perplexity are true, the company is violating many of those principles.

“By default, AI companies should assume they have no right to take and reuse publishers' content without permission,” Kint says. If Perplexity is skirting terms of service or robots.txt, he adds, "the red alarms should be going off that something improper is going on.”

Amazon Is Investigating Perplexity Over Claims of Scraping Abuse

Gutted of civil rights protections by Democrats to woo pro-business Republicans, the American Privacy Rights Act was pulled from a key congressional hearing—and appears unlikely to receive a full vote.

Dozens of civil rights organizations had been urging Democrats (some of whom had puzzlingly signed off on those changes) to sink the bill, arguing that the changes were both “immensely significant and unacceptable.”

The new text, engineered to appease conservative lobbyists representing the interests of big business, omitted, for instance, a key section referencing “civil rights.” The deleted section aimed to prevent businesses from trafficking in people’s data “in a manner that discriminates in or otherwise makes unavailable the equal enjoyment of goods or services on the basis of race, color, religion, national origin, sex, or disability.” For reasons that at this stage are above obvious, GOP lawmakers are firmly opposed to such language.

Deleting sections of a bill holding companies accountable for making data-driven decisions that could lead to discrimination in housing, employment, health care, and the like spurred a strong response from civil society organizations including the NAACP, the Japanese American Citizens League, the Autistic Self Advocacy Network, and Asian Americans Advancing Justice, among dozens of others.

In a letter this week to E&C Democrats, obtained by WIRED, the groups wrote: “Privacy rights and civil rights are no longer separate concepts—they are inextricably bound together and must be protected. Abuse of our data is no longer limited to targeted advertising or data breaches. Instead, our data are used in decisions about who gets a mortgage, who gets into which schools, and who gets hired—and who does not.”

But the cuts did not end there. The most recent version of the ARPA noticeably excluded language designed to grant users the power to opt-out before companies could use algorithms to “facilitate a consequential decision” using an individual’s personal data. At the same time, language that would have imposed a duty on companies to examine, or audit, the impacts of their own algorithms on users was likewise erased.

Both of these provisions contained generous “pro-business” caveats. For instance, users would be able to opt out of algorithmic decisionmaking only if doing so wasn’t “prohibitively costly” or “demonstrably impracticable due to technological limitations.” Similarly, companies could have limited the public’s knowledge about the results of any audits by simply hiring an independent assessor to complete the task rather than doing so internally.

“Prior versions of APRA required companies that developed or used AI for making automated decisions about people in certain important areas like employment, housing, and credit to be transparent about those systems and to allow people to opt out of that automated decisionmaking,” says Eric Null, codirector of the privacy and data project at the Center for Democracy & Technology, a digital rights nonprofit. “Without those provisions, people can and will be subject to AI that makes or contributes to important, life-changing decisions about them, and they will have little to no way to protect themselves.”

Digital rights groups such as Access Now, Demand Progress, and Free Press Action joined in to pressure Democrats not to accept these changes in stride, arguing that “a privacy bill that does not include civil rights protections will not meaningfully protect us from the most serious abuses of our data,” and that the changes were imposed “without prior stakeholder consultation and without studying the impact to the bill’s ability to address data-driven discrimination.”

WIRED had reached out on Wednesday to 23 Democrats currently serving on the E&C to get a response to the demands of these groups. A single lawmaker responded:

“I already had concerns with the American Privacy Rights Act,” US representative Nanette Barragán said, pointing to language in the bill that could arguably undermine stronger data privacy protections already implemented by her home state of California. “The latest draft only deepens my concerns about the bill because critical civil rights provisions have been removed from the proposal.”

In a statement after Thursday’s cancellation, the E&C’s ranking Democrat, Frank Pallone, Jr., blasted GOP leaders for interfering with the committee’s process while at the same time extending his gratitude to the committtee’s Republican chair, Cathy McMorris Rodgers, lauding her dedication to “giving Americans back control of their data.”

“We’re not giving up,” adds Pallone, declaring he and his colleagues are the only ones in Congress with the guts to “take on Big Tech on behalf of the American people.”

Surprise! The Latest ‘Comprehensive’ US Privacy Bill Is Doomed

A custom platform developed by SITU Research aided the International Criminal Court’s prosecution in a war crimes trial for the first time. It could change how justice is enacted on an international scale.

In 2012, the al-Qaeda–linked jihadist group Ansar Eddine swept through Timbuktu, taking control of the Malian city. As part of the group’s imposition of sharia law over the residents, it created an Islamic police force, led by Al Hassan Ag Abdoul Aziz. Ansar Eddine controlled the city until 2013, when a military campaign led by France drove the group out.

During his tenure with the police force, Al Hassan helped carry out a campaign of “forced marriages” and floggings of those accused of violating the strict religious rules imposed by Ansar Eddine. He faced 13 counts of crimes against humanity before the International Criminal Court at the Hague. Earlier today, the ICC convicted Al Hassan of war crimes and crimes against humanity, including torture and cruel treatment, for which he could face life in prison.

As part of the case, prosecutors employed a unique strategy of presenting evidence—a bespoke immersive replica of the city that includes visual evidence of where the alleged crimes took place. Created by SITU Research, a visual investigations firm, the platform used to prosecute Al Hassan represents a watershed moment in how evidence of war crimes is presented and the kinds of evidence the ICC might consider in future rulings.

Established in 2002, the ICC is relatively young. Alexa Koenig, a law professor at UC Berkeley, who helped set up the Technology Advisory Board at the ICC, says that the impetus for incorporating new technologies into evidence came as a response to an evaluation from the court that cases relied too heavily on witness testimony.

“Judges were saying that the prosecution had a problem,” says Koenig. “That essentially they were over-relying on the stories of survivors without bringing in the corroborating information needed to meet the evidentiary standards to allow these cases to go forward.”

As more and more evidence of war crimes became digital—through social media content or videos and photos uploaded online—Koenig says, the court has allowed this material to be submitted as evidence, but other tools, like the platform created by SITU Research, are still a nascent way to present evidence. And because cases can take much longer to move through the ICC than another kind of court, Koenig says, the ICC has only begun to reckon with these kinds of questions.

“As people are preserving evidence on places like Facebook or WhatsApp or YouTube, we're going, How do we preserve this in a way that preserves its values for our purposes?” Koenig says. “But also how do we make the content more impactful?”

SITU Research created a similar platform for the ICC prosecution of Ahmad Al Faqi Al Mahdi, also a part of Ansar Eddine, who was accused of destroying sites of cultural heritage in Timbuktu. But al Mahdi pleaded guilty, so the court never had the opportunity to evaluate the platform as a piece of evidence. Al Mahdi was sentenced to nine years in prison in 2016.

“It's kind of challenging to get new tech integrated and admitted as evidence,” says Brad Samuels, founding partner at SITU Research. “And often the technology that's in the world isn't necessarily finding its way into the courtroom quickly.”

The platform allows users to virtually walk through the city of Timbuktu, as well as see it from above. Nested within the virtual re-creation of the city are pieces of video, photo, and audio evidence, showing where and when an event allegedly occurred. For instance, users can explore down a street and encounter an item labeled to a corresponding piece of evidence, click on it, and see a video of a particular moment that shows who was allegedly present and what occurred.

SITU’s platform combines drone footage, satellite imagery, laser scans, and other forms of visual evidence that, Samuels says, allows the judges to see the evidence over time and space in a way that would not otherwise be possible. This is particularly important in cases of crimes against humanity, which require the crimes to be systematic and widespread as opposed to a single, isolated incident. And he says there’s more value to the platform than just helping the judges organize visual evidence that supports witness testimony.

“In the case of Timbuktu, the security situation in Mali has sort of gotten worse and worse. And the judges definitely were not able to go there in person,” Samuels says. For instance, the platform includes the interior of a former bank that functioned as the headquarters of Islamic Police, and another site where someone was flogged.

“It can also become a tool for interviewing witnesses about what they experienced,” Koenig says of SITU’s platform. “If you think about some of the language barriers and cultural barriers, not having a touchpoint like that, have them walk you through what happened to them can be almost guesswork at times of trying to piece together their story, particularly when there's a lot of trauma involved and memories are fragmented or recollections are fragmented.”

Sarah Zarmsky, assistant lecturer at Essex Law School, says, however, that often the defense teams for the accused lack the same resources offered to the prosecution.

“This platform was designed for the prosecution,” she says. “So, while SITU Research does great work, the defense also isn't coming in with that same expertise. And so there are a lot of fair-trial implications there.”

In an article published in 2023, ICC chief prosecutor Karim Khan noted that “the defense sought full access to the virtual platform, and required (and were provided) training and guidance on its use, so that they could also deploy it as needed.”

In the case of the ICC, where the court allows almost all evidence to be submitted and then decides afterward how much weight to assign each piece of evidence, that can mean an overburdened defense team may not have the time to sift through each piece of evidence. For something novel like the platform created by SITU, Zarmsky says, it’s unclear how much weight the court might put on it when it comes down to making a ruling.

But Khan asserted that the “Trial Chamber’s acceptance of the platform shows that the law and procedure of the ICC is ready to accommodate forward-thinking approaches to the presentation of evidence.”

War Crime Prosecutions Enter a New Digital Age

WikiLeaks founder Julian Assange has agreed to plead guilty to one count of espionage in US court on Wednesday, ending a years-long legal battle between the US government and a controversial publisher.

United States prosecutors have secured a deal with WikiLeaks founder Julian Assange requiring the long-embattled publisher to plead guilty to one count of espionage for his role in making public classified documents concerning the US wars in Iraq and Afghanistan.

The agreement, which follows more than a decade of efforts by Assange, 52, to avoid extradition from the United Kingdom, would draw to a close one of the longest-running national security investigations in US history. The deal was first disclosed in court documents made public in the UK.

Assange and his legal team, which have denied the accusations levied by the US, could not be immediately reached for comment.

“Julian Assange is free,” WikiLeaks wrote in a statement posted to X. “He left Belmarsh maximum security prison on the morning of 24 June, after having spent 1901 days there.”

A letter US prosecutors filed in the US District Court for the Northern Mariana Islands on Monday indicates that Assange will enter his guilty plea at a Wednesday hearing in Sapian, the island territory’s capital, having refused to travel to the continental US. He is then expected to return to his home country of Australia, having already served the expected 62-month sentence in London prison.

The case against Assange centers around the publishing of more than 750,000 stolen US documents by WikiLeaks between 2009 and 2011. It has drawn enormous attention for its clear implications on press freedoms internationally. Organizations such as the Committee to Protect Journalists in the US have for years warned the case could severely imperil the ability of journalists to obtain and publish classified information—even though the nation’s highest court has long recognized the right of journalists to do so.

Ahead of the 2016 US presidential election between Hillary Clinton and Donald Trump, WikiLeaks published a trove of emails stolen from the Democratic National Committee. The leak, which embarrassed the DNC and won Assange praise from right-wing figures, was later revealed to be the work of notorious Russian hacking groups known as Cozy Bear and Fancy Bear, both affiliated with Moscow’s GRU military intelligence agency.

US prosecutors initially charged Assange with a single count under the Computer Fraud and Abuse Act for allegedly conspiring with Chelsea Manning, who provided WikiLeaks with the trove of classified material related to the wars in Iraq and Afghanistan, to gain unauthorized access to government computers. Prosecutors later added an additional 17 charges under the Espionage Act—a move widely condemned as an attack on the free press.

Assange, forcibly removed from the Ecuadorian embassy in London in 2019 after seven years asylum, has been held in Belmarsh prison in London pending the outcome of his extradition hearings, which were delayed repeatedly over the course of the Covid-19 pandemic. His attorneys argued that due to his deteriorating mental health, extradition to the US would increase the likelihood of suicide.

US prosecutors secured, on appeal, permission to extradite the award-winning journalist, who married his longtime partner, Stella Moris, while in jail in 2022, by offering UK courts a slate of written assurances. Among other concessions, the US promised not to subject Assange to “special administration measures,” a term referring to the practice of wiretapping certain defendants’ phone calls citing national security concerns.

“This period of our lives, I’m confident now, has come to an end,” said Moris—now Assange—in a video prerecorded last week. “I think by this time next week, Julian will be free.”

Kristinn Hrafnsson, WikiLeaks editor in chief, said in the same video captured outside Belmarsh that he hoped to see Assange for the last time inside its walls. “If you’re seeing this, it means he is out.”

The Julian Assange Saga Is Finally Over

How accurate are gunshot detection systems, really? For years, it's been a secret, but new reports from San Jose and NYC show these systems have operated well below their advertised accuracy rates.

Liz González’s neighborhood in East San Jose can be loud. Some of her neighbors apparently want the whole block to hear their cars, others like to light fireworks for every occasion, and occasionally there are gunshots.

In February 2023, San Jose began piloting AI-powered gunshot detection technology from the company Flock Safety in several sections of the city, including Gonzalez’s neighborhood. During the first four months of the pilot, Flock’s gunshot detection system alerted police to 123 shooting incidents. But new data released by San Jose’s Digital Privacy Office shows that only 50 percent of those alerts were actually confirmed to be gunfire, while 34 percent of them were confirmed false positives, meaning the Flock Safety system incorrectly identified other sounds—such as fireworks, construction, or cars backfiring—as shooting incidents. After Flock recalibrated its sensors in July 2023, 81 percent of alerts were confirmed gunshots, 7 percent were false alarms, and 12 percent could not be determined one way or the other.

For two decades, cities around the country have used automated gunshot detection systems to quickly dispatch police to the scenes of suspected shootings. But reliable data about the accuracy of the systems and how frequently they raise false alarms has been difficult, if not impossible, for the public to find. San Jose, which has taken a leading role in defining responsible government use of AI systems, appears to be the only city that requires its police department to disclose accuracy data for its gunshot detection system. The report it released on May 31 marks the first time it has published that information.

The false-positive rate is of particular concern to communities of color, some of whom fear that gunshot detection systems are unnecessarily sending police into neighborhoods expecting gunfire. Nonwhite Americans are more often subjected to surveillance by the systems and are disproportionately killed in interactions with police. “For us, any interaction with police is a potentially dangerous one,” says Gonzalez, an organizer with Silicon Valley De-Bug, a community advocacy group based in San Jose.

San Jose did not attempt to quantify how many shooting incidents in the covered area the Flock System failed to detect, also known as the false-negative rate. However, the report says that “it is clear the system is not detecting all gunshots the department would expect.”

Flock Safety says its Raven gunshot detection system is 90 percent accurate. SoundThinking, which sells the ShotSpotter system, is the most popular gunshot detection technology on the market. It claims a 97 percent accuracy rate. But the data from San Jose and a handful of other communities that used the technologies suggest the systems—which use computer algorithms, and in SoundThinking’s case, human reviewers, to determine whether the sounds captured by their sensors are gunshots—may be less reliable than advertised.

Last year, journalists with CU-CitizensAccess obtained data from Champaign, Illinois, showing that only 8 percent of the 64 alerts generated by the city’s Raven system over a six-month period could be confirmed as gunfire. In 2021, the Chicago Office of Inspector General reported that over a 17-month period only 9 percent of the 41,830 alerts with dispositions that were generated by the city’s ShotSpotter system could be connected to evidence of a gun-related crime. SoundThinking has criticized the Chicago OIG report, saying it relied on “incomplete and irreconcilable data.”

This week, New York City’s comptroller published a similar audit of the city’s ShotSpotter system showing that only 13 percent of the alerts the system generated over an eight-month period could be confirmed as gunfire. The auditors noted that while the NYPD has the information necessary to publish data about ShotSpotter’s accuracy, it does not do so. They described the department’s accountability measures as “inadequate” and “not sufficient to demonstrate the effectiveness of the tool.”

Champaign and Chicago have since canceled their contracts with Flock Safety and SoundThinking, respectively.

“Raven is over 90 percent accurate at detecting gunshots with around the same accuracy percentage at detecting fireworks,” Josh Thomas, Flock Safety senior vice president of policy and communications, tells WIRED in a statement. He adds that this is “based on measurements across the board,” meaning data from all Flock's customers. “And critically, Raven alerts officers to gun violence incidents they never would have been aware of. In the San Jose report, for example, of the 111 true positive gunshot alerts, SJPD states that only 6 percent were called in to 911.”

Eric Piza, a professor of criminology at Northeastern University, has conducted some of the most thorough studies available on gunshot detection systems. In a recent study of shooting incidents in Chicago and Kansas City, Missouri, his team’s analysis showed that police responded faster to shooting incidents, stopped their vehicles closer to the scene of shootings, and collected more ballistic evidence when responding to automated gunshot alerts compared to 911 calls. However, there was no reduction in gun-related crimes, and police were no more likely to solve gun crimes in areas with gunshot sensors than in areas without them. That study only examined confirmed shootings; it did not include false-positive incidents where the systems incorrectly identified gunfire.

In another study in Kansas City, Piza found that shots-fired reports in areas with gunshot sensors were 15 percent more likely to be classified as unfounded compared to shots-fired reports in areas without the systems, where police would have relied on calls to 911 and other reporting methods.

“If you look at the different goals of the system, research shows that \[gunshot detection technology\] typically tends to result in quicker police response times,” Piza says. “But research consistently has shown that gun violence victimization doesn’t reduce after gunshot detection technology has been introduced.”

The New York City comptroller recommended the NYPD not renew its current $22 million contract with SoundThinking without first conducting a more thorough performance evaluation. In its response to the audit, the NYPD wrote that “non-renewal of ShotSpotter services may endanger the public.”

In its report, San Jose’s Digital Privacy Office recommended that the police department continue looking for ways to improve accuracy if it intends to keep using the Raven system.

Pointing to the report’s finding that only 6 percent of the confirmed gunshots detected by the system were reported to police via 911 calls or other means, police spokesperson Sergeant Jorge Garibay tells WIRED the SJPD will continue to use the technology. “The system is still proving useful in providing supplementary evidence for various violent gun crimes,” he says. “The hope is to solve more crime and increase apprehension efforts desirably leading to a reduction in gun violence.”

The department’s media relations division could not immediately locate information about how much SJPD pays Flock Safety for the Raven system or how long its contract is for.

Darcie Green, a community health advocate and member of San Jose’s former Reimagining Public Safety Community Advisory Group, says that disclosing data about the accuracy of the system is a good step, but the city also needs to examine whether the technology is actually making the city safer or whether the money and human resources devoted to responding to real and false gunshot alerts could be better invested elsewhere.

“Obviously we want to reduce gun violence, that’s everybody’s goal,” she says. “For these dollars that we’re spending on enforcement and punishment, we should be spending three-, four-, fivefold on programming and improvement. There are so many people for whom calling the police is not a solution.”

Gonzalez, the East San Jose community organizer, says the lack of 911 calls about shootings is a reflection of how little she and others in the community trust police responses to their neighborhood.

“There was a time when a helicopter was searching for somebody in our neighborhood, and I could swear that they were in my backyard, but I didn’t want to call the police because potentially they could shoot me or my family mistaking me for that person,” she says. “I would rather have the person in my backyard than the police looking for them.”

_Updated 6/25/2024, 3:55 pm ET: Added additional detail about how Flock Safety calculates its accuracy levels._

The Mystery of AI Gunshot-Detection Accuracy Is Finally Unraveling

The most notorious deepfake sexual abuse website is hosting altered videos originally published as part of the GirlsDoPorn operation. Experts say this new low is only the beginning.

_This article contains descriptions of sex trafficking and abuse. Discretion is advised._

For years, nonconsensual deepfake pornography has been used to harass, silence, shame, and abuse women. Celebrities and influencers have their faces implanted into existing adult videos; men have used the technology to place “friends” into explicit videos; and boys have allegedly created “nude” images of their female classmates. However, among the ever growing harassment and abuse, deepfake creators have now, arguably, hit a new low: using videos of sex trafficking victims as the basis of the nonconsensual videos.

Over the past two months, an account on the largest deepfake sexual abuse website has posted 12 celebrity videos that are based on footage from GirlsDoPorn, a now-defunct sex trafficking operation that the US Department of Justice says its operators used to conspire and commit sex trafficking through “force, fraud, and coercion,” tricking five women—and allegedly hundreds more— into making sex videos that were subsequently posted online.

The dozen videos—which ran up to 21 minutes long and racked up tens of thousands of views before they were taken down following WIRED’s inquiry—used footage originally posted to the GirlsDoPorn website and had celebrity faces added using artificial intelligence. It appears that a startup’s face-swapping tool may have been abused to transform the videos into deepfakes, according to watermarks on the footage.

As deepfake technology has become increasingly capable of creating realistic imagery and easier to use, hundreds of websites and apps designed to create or host deepfake sexual abuse have appeared. Laws to limit the use of these tools and protect those targeted are lagging, even as the malicious use of AI has evolved to not just create new victims but to revictimize survivors.

From 2012 to 2019, the DOJ says, the creators of GirlsDoPorn worked by recruiting young women, using ads posted on Craigslist, for what they claimed were clothed modeling photo shoots. When the women responded, they were told the ads were for pornographic videos, and they were pressured into taking part, according to various lawsuits and survivor testimonies.

The individuals behind the scheme, according to the DOJ, told the women the videos would only be sold on DVDs outside of the United States and the footage would not be posted to the internet. Instead, they then posted short videoclips to online platforms, such as Pornhub, and full-length versions to their GirlsDoPorn website. The videos have circulated online since.

Multiple legal proceedings against the creators of GirlsDoPorn, people affiliated with the website, and Pornhub's parent company are ongoing or have been completed—with criminal charges being issued against several GirlsDoPorn organizers in October 2019.

Since then, 22 survivors have been awarded nearly $13 million in damages, and 402 victims were given copyright ownership of videos featuring them, making it easier to scrub them from the web. At the end of 2023, Pornhub’s parent company, Aylo Holdings, agreed to pay damages to women impacted by the sharing of the GirlsDoPorn videos.

Among those charged, Ruben Andre Garcia, a GirlsDoPorn producer and recruiter, was sentenced to 20 years in prison; Matthew Isaac Wolfe, who admitted to having a “wide range of responsibilities” at GirlsDoPorn, according to the DOJ, was sentenced to 14 years; cameraman Theodore Wilfred Gyi was sentenced to four years; and GirlsDoPorn bookkeeper Valorie Moser pleaded guilty to one charge of conspiracy to commit sex trafficking and is awaiting sentencing. Finally, in March this year, the alleged GirlsDoPorn mastermind, Michael Pratt, was extradited from Spain to the US to face charges linked to the operation. He has pleaded not guilty. In total, those involved in GirlsDoPorn have been ordered to pay more than $35 million in restitution.

Brian Holm, a managing attorney at the Holm Law Group and a longtime civil attorney for GirlsDoPorn survivors, confirmed that the videos posted to the deepfake sexual abuse website were originally from GirlsDoPorn. These include, Holm says, survivors who have been involved in the legal cases against GirlsDoPorn or against Pornhub.

“It’s a real double whammy for trafficking victims to see their videos used like this,” says Holm, adding that the videos are the tip of the iceberg. “From what I’ve seen on that site, I think there's 10 times the amount you sent me that I’ve seen on there.”

The 12 videos posted by the account seen by WIRED have received up to 15,000 views each, and several have a ‘girlsdoporn.com’ watermark on the footage. The account that posted the videos has a version of “GirlsDoPorn” as its username and included the sex trafficking site in the title of the videos.

WIRED is not naming the deepfake abuse website due to its role in spreading abusive content or the celebrities featured in the videos. The website is the largest website of its kind—hosting tens of thousands of videos and receiving millions of visitors. In April, the website blocked visitors from the UK after lawmakers in the country announced plans to make it a criminal offense to create nonconsensual explicit deepfakes.

“These creators of sexually explicit deepfakes have no regard whatsoever for the women and girls who are victims of sex trafficking and now being further abused through this deepfake sexual abuse,” says Clare McGlynn, a professor of law at Durham University, who works to counter image-based abuse.

“This website is actively choosing to share recordings of actual sexual assaults,” McGlynn says. “These are heinous acts, deliberately and knowingly causing life-shattering and life-threatening harms. The drive for profit, for fueling the trade in nonconsensual porn, knows no bounds. This shows a contempt for the rights of women and girls.”

Neither the account posting the deepfake GirlsDoPorn videos nor the site’s anonymous administrator’s replied to questions from WIRED.

At the end of March, another user on the website asked whether the GirlsDoPorn footage was allowed, saying it made them “feel sick.” They suggested some people may not know the history of GirlsDoPorn but pointed out: “This … one user clearly does tho, with branding themselves with a rape website.” A moderator replied saying they were not going to remove the videos but said if there was a “list” of videos confirmed to include sex trafficking victims they would notify the website’s administrators to take them down.

As well as the GirlsDoPorn watermark, the 12 videos hosted on the deepfake website also have a watermark of US company Akool, which offers generative AI services, including a “face-swapping” tool, to marketing professionals.

Jiajun Lu, the founder and CEO of Akool, says the company’s terms of service prohibit copyright violations, sexual and violent content, and it frequently bans accounts. “They are strictly forbidden on our platform, and we have multiple approaches to prevent it from happening, and more solutions are coming up soon,” he says. “We changed our watermarks a few months ago, and these videos might not come from our website.”

The CEO says that the company is working on new safety tools as it increases the security of its website: For instance, it is considering verifying the identity of people trying to use its face-swapping tool using face recognition, and initially the team is building more prominent warning notices about prohibited use cases.

Since 2016, Charles DeBarber, director of operations at technology firm Phoenix AI, has been undertaking the arduous task of discovering and trying to remove content from GirlsDoPorn that has been posted online. As part of Phoenix AI, this work—including building tools to automate online searches and using face recognition—has helped to remove around 60 GirlsDoPorn survivors’ footage from the web, DeBarber says.

“Not only does it open wounds of survivors already, but it's making new survivors and making new victims,” DeBarber, who has worked with GirlsDoPorn survivors for years, says of the deepfakes. The deepfake abuse website, DeBarber says, doesn’t take anything down when requests are made to it, and it is hosted by a company registered in the Seychelles.

Since deepfake sexual abuse videos appeared more than half a decade ago, those targeted by them have struggled to get them removed from the web, and US lawmakers have been slow to regulate or criminalize the abusive behavior. In many instances, including in wider cases of image-based abuse, people trying to remove abusive content from the web have had to resort to complaints made under the Digital Millennium Copyright Act. According to DeBarber, it’s not enough.

“Our laws are written in a way that protects intellectual property but not survivors, and it’s very demeaning for a lot of our survivors,” he says. “These are people that were treated like a commodity, and now we have to use intellectual property law to get it down.”

Deepfake Creators Are Revictimizing GirlsDoPorn Sex Trafficking Survivors

With cyberattacks increasingly targeting health care providers, an arduous bureaucratic process meant to address legal risk is keeping hospitals offline longer, potentially risking lives.

Crippling ransomware attacks against hospitals and health care providers are on the rise. These ruthless cyberattacks can take medical systems offline for weeks—canceling appointments and surgeries and causing harm to patients. Doctors and nurses are plunged into crisis situations where they resort to using pen and paper, while IT staff work to make systems safe and bring them back online. The recovery can be long-lasting and brutal.

Health care professionals, lawyers, and cybersecurity experts tell WIRED that amid the chaos caused by criminal hackers, a little-known bureaucratic process can slow down hospitals and medical providers getting their systems working again.

The red tape involves organizations hit by ransomware sending detailed “assurance” or “attestation” letters to companies that they connect their systems or software with. These letters are designed to convince organizations that it is safe to reconnect after the ransomware attack, but they can add extra pressure to those already dealing with physically and mentally draining recovery operations.

The letters aren’t required by any law and are not unique to medical organizations impacted by ransomware attacks, but experts say in situations where lives are at risk, more efficient processes should be considered. Assurance letters seen by WIRED contain up to 40 individual questions about cyberattacks and include detailed requests about how events unfolded, steps taken to respond, and any evidence that may have been gathered.

“Negotiating with hundreds of vendors each with their own unique set of requirements to reconnect was an arduous and time-consuming process,” says Sean Fitzpatrick, the vice president of external communications at Ascension, a network of 140 hospitals and thousands of affiliated providers across 19 states, which was hit by ransomware in May. Ascension now has more than 95 percent of its suppliers reconnected or in the process of reconnecting, Fitzpatrick says, and has attempted to be as transparent as possible with its recovery.

Meanwhile, Shane Thielman, the chief information officer at Scripps Health, a San Diego health care provider with more than 70 hospitals and clinics, says it was required to produce around 30 vendor letters when it was hit by Conti malware in May 2021 for clinical, business, and administrative software partners. “Initially, there were several vendors that did not accept the assurance letter and requested additional technical documentation and information,” Thielman says. “However, this was limited and ultimately did not result in a delay to restoring access to systems at Scripps.”

**Lawyers Enter the Chat**

Hospitals and other health care organizations are hugely complex beasts—they can use dozens of pieces of software from different companies, providing everything from electronic health care records to staff shift schedules. Cyberattacks that take medical services offline can impact a hospital’s own internal systems and networks and those belonging to third-party suppliers. An attack against a company that provides software to hundreds or thousands of organizations can have widespread ramifications.

It doesn’t take long to see how ransomware, which is most often launched by Russia-based cyberciminals, can have a dangerous impact on medical care. More than 1,100 surgeries and 2,194 appointments have been canceled at London hospitals this month after pathology services provider Synnovis was hit by ransomware. Some cancer appointments have been canceled, blood testing reduced, and more than 50 organs needed to be diverted for use at other hospitals, figures from the UK’s National Health Service show.

“I can tell you with complete confidence that ransomware attacks harm patients,” says Hannah Neprash, an associate professor of health policy at the University of Minnesota, who has researched the impact of ransomware attacks on US hospitals and concluded they result in higher mortality rates. “If you are a patient who has the misfortune to be admitted to a hospital when that hospital goes through a ransomware attack, the likelihood that you're going to walk out the doors goes down,” Neprash says. “The longer the disruption, the worse the health outcomes.”

In the hours and days immediately after ransomware attacks, it’s common for companies who have software connected to the targeted organization to pull their services. This can include everything from disconnecting medical records to refusing to email a cyberattack victim. This is where so-called assurance letters come in.

“We’ve really seen the demand for these letters increase over the past few years as breaches have become much more litigious—from class actions lawyers chasing settlements to lawsuits between businesses,” says Chris Cwalina, the global head of cybersecurity and privacy at law firm Norton Rose Fulbright.

Cwalina says he is unsure where and when the practice of sending assurance letters started but says it is likely it began with lawyers or security professionals who misunderstood legal requirements or the risks they are trying to prevent. “There is no legal requirement to request or obtain an attestation before systems can be reconnected,” Cwalina says.

These assurance and attestation letters are often compiled with the support of specialist cybersecurity companies that are employed to respond to incidents. What can be reconnected and when will vary depending on the specific details of each attack.

But much of the decisionmaking comes down to risk—or at least perceived risk. Charles Carmakal, the chief technology officer of Google-owned cybersecurity firm Mandiant, says companies will be worried that cybercriminals could move “laterally” between the victim and their systems. Companies want to know a system is clean and the attackers have been removed from the systems, Carmakal says.

“I understand the rationale behind the assurance process. What I would say is that people do need to really consider what is the risk associated with the level of connectivity between two parties, and sometimes people tend to default to the most restrictive path,” Carmakal says. For instance, it is rare that Mandiant sees wormable ransomware moving from one victim to another, he says.

“Vendors were interested to know that independent, outside cybersecurity experts were engaged with Scripps technical teams and verification that malware was contained and remediated with reasonable best efforts,” Thielman, the CIO of Scripps Heath, says. For Ascension, Fitzpatrick says, the company also held one-on-one calls with vendors and hosted eight webinars where it provided updates. It has also shared indicators of compromise—the traces left by attackers in its systems—with health organizations and the US Cybersecurity and Infrastructure Security Agency (CISA).

**Third-Party Doctrine**

Cybercriminals have become more brazen with attacks against hospitals and medical organizations in recent years; in one case, the Lockbit ransomware gang claimed it had rules against attacking hospitals but hit more than 100. Often these sort of attacks directly impact private sector companies that provide services to public infrastructure or medical organizations.

“If you look plausibly at the threat picture in the years ahead, disruption to public services and public activity caused by \[cybercrime\] activity that affects the private sector is probably something that's going to happen more and more,” says Ciaran Martin, a professor at the University of Oxford and the former head of the UK’s National Cyber Security Centre. In these instances, Martin suggests, there may be questions around whether governments have, or need, powers to direct private firms to respond in certain ways.

John Riggi, national adviser for cybersecurity and risk at the American Hospital Association, says there could be more federal support for health care providers. “Some of the resources, for instance, that CISA dedicates to defending federal civilian networks could also be devoted and expanded to help defend critical infrastructure, like health care.”

When it comes to the need for assurance and attestation letters, Norton Rose Fulbright’s Cwalina says much of the information requested in these scenarios could be received through directly speaking to people. “The piece of paper is ultimately only obtained after lawyers review it and insert caveats and protections—which means the attestation will actually say very little to give an organization confidence to reconnect and certainly nothing that couldn’t be obtained through other means,” Cwalina says. Scripps Health’s Thielman recommends establishing communications with software providers as soon as possible.

“We need a situation in which somebody, be that a incident response company or CISA, is able to sound an ‘all clear’ signal after which vendors can reconnect immediately,” says Brett Callow, a threat analyst at Emsisoft. A third-party approval could speed up the process, he says. “The primary concern here has to be patient safety. We need to get hospitals up and running in a safe manner, in the shortest possible time. Masses of red tape just aren’t conducive to that.”

Red Tape Is Making Hospital Ransomware Attacks Worse

Plus: Alleged Apple source code leaks online, cybercrime group Scattered Spider's alleged kingpin gets arrested, and more.

The rolling series of breaches targeting customers of cloud platform Snowflake appears to be a supply chain attack wrapped in another supply chain attack. A hacker who claims to have been involved in the attacks tells WIRED that the hackers, known as ShinyHunter, stole victims’ Snowflake credentials by first breaching an employee of a third-party contractor. (The contractor, however, says it does not believe it was involved.)

Ultimately, the breach of the Snowflake customer accounts, which include Ticketmaster, banking firm Santander, and potentially more than 160 other companies, was possible because their Snowflake accounts did not have multifactor authentication enabled.

Antivirus giant Kaspersky’s worst nightmare has finally come true: The United States government announced on Thursday that it is banning the sale of its software to new customers in the US over alleged Russian national security threats. (Kaspersky has challenged the Biden administration’s claims.) Existing customers, meanwhile, will be banned from downloading Kaspersky software updates after September 29. What could go wrong?

Perplexity AI, an artificial-intelligence-powered search startup, says it’s already valued at a billion dollars. But a WIRED investigation published this week found that its secret sauce has a pungent ingredient: bullshit.

Beyond “hallucinating” details generated by its chatbot, WIRED found that the AI tool appears to be ignoring the Robots Exclusion Protocol—a standard web tool used to prevent scraping—on sites owned by WIRED’s parent company, Condé Nast, and other publications, seemingly allowing it to scrape articles despite the internet equivalent of a “Do Not Enter” sign hanging on WIRED and other Condé Nast sites. Perplexity’s chatbot later plagiarized that same article when prompted.

People traveling through some of the largest train stations in the United Kingdom secretly had their faces scanned by Amazon’s face-recognition tools, according to documents obtained by WIRED. The technology, which was used as part of a trial run, predicted travelers’ various attributes, including gender, age, and likely emotions. The surveillance, which one privacy advocate called “concerning,” could potentially be used for serving advertisements.

Finally, we detailed the rise of robot “dogs” used by militaries, explained what would happen if China invaded Taiwan, and got into the nitty-gritty of the boring-sounding but serious work of spotting the billion-dollar scam tactic known as business email compromise.

That’s not all. Each week, we round up the security and privacy news we didn’t cover in depth ourselves. Click the headlines to read the full stories. And stay safe out there.

For months, ransomware gangs have rampaged across the health care industry, with ruthless attacks targeting Change Healthcare’s national payment network for more than a thousand health care providers, Ascension Healthcare’s 140 hospitals, and dozens of other victims in the medical field. Now that hacking epidemic is crystallizing into yet another catastrophic hospital hack—one that has resulted in the data of 300 million UK patient records leaking online.

Synnovis, a joint-venture medical testing company partially owned by the UK’s National Health Service, has for weeks been battling and negotiating with the Russia-linked ransomware group Qilin, which has deeply disrupted its services in an attempt to extort the company. The result has been well over a thousand postponed operations and thousands more postponed outpatient appointments across multiple UK hospitals. Ambulances have been diverted from the affected hospitals, potentially causing delays in lifesaving care. They’ve even had to ask for new urgent donations of O-type blood, as testing disruptions have prevented other types from being used in patients’ blood transfusions.

Source

Wired