Three states are suing to block security rules for water facilities. If they win, it may open the floodgates for challenges to other cyber rules.

The White House doesn’t have the same control over the EPA, which is an independent agency, but Greene says that from what he saw, the agency tried to collaborate with the water sector.

The NSC did not respond to a request for comment about the EPA lawsuit and its possible effects on the administration’s agenda. The EPA declined to comment because the litigation is pending.

A Legal Fight on Multiple Fronts

The Republican attorneys general challenging the EPA directive make several claims. They say the agency failed to follow the proper procedure for issuing a regulation. They allege that the EPA exceeded its authority under the Safe Drinking Water Act and subsequent legislation. And they argue that, by requiring state water regulators to fold cybersecurity into their inspections, the federal government is usurping states’ sovereign authority to regulate water facilities and unconstitutionally burdening them with new work.

Michael Blumenthal, an environmental regulation lawyer at McGlinchey Stafford, says the EPA did appear to have violated the Administrative Procedure Act by issuing its directive to states as a reinterpretation of existing guidance about states’ responsibilities to conduct “sanitary surveys” of water facilities, thus sidestepping the public comment process.

Peggy Otum, a partner at WilmerHale who leads the law firm’s environment practice, says the state-sovereignty argument reflects a broader debate about how much the federal government—and the EPA in particular—can burden states with new mandates. “‘Who's gonna pay for it?’ is the main question,” Otum says.

Greene was skeptical of this argument. The White House is aware of the water sector’s funding issues, he says, but that’s not a good enough reason to refrain from mandating better security.

Open for Interpretation

But the most consequential argument in the case concerns whether the EPA’s regulatory authority for the water sector even extends to cybersecurity. Blumenthal says the Safe Drinking Water Act “does not give them the authority to fold in cybersecurity.”

The EPA derived its authority from newly reinterpreted definitions of key terms in its guidance to states, but Blumenthal says that approach was invalid and would allow mandates that were “never contemplated to begin with.”

Greene argues that laws like the Safe Drinking Water Act, while enacted before cyber threats gained prominence, were clearly intended to let the EPA protect vital resources against all manner of dangers. “It would be an overly literal reading of the intent of these \[laws\] to say, ‘They didn’t think about cybersecurity, so you can't cover it,’” Greene says. “That's like saying, ‘The colonial armies didn’t think about air assets.’”

Courts have historically deferred to agencies in lawsuits over the interpretation of their core statutes, but this principle, known as _Chevron_ deference, “is hanging on by a thread” at the US Supreme Court, Otum says.

“Everyone's Sniffing Around”

The EPA lawsuit looms large as a potential stumbling block for the Biden administration’s new national cyber strategy, which describes critical infrastructure regulation as a national security imperative. Other regulators “are going to watch this case very closely to see what happens,” Blumenthal says.

The Department of Health and Human Services is working on cyber rules for hospitals, which, like water facilities, are heavily regulated by states. The Federal Communications Commission (FCC) is preparing rules to secure the Emergency Alert System, a critical tool for state and local authorities. And the Federal Trade Commission (FTC) is updating its security regulations and sharpening its oversight of data breach disclosures.

A Republican-Led Lawsuit Threatens Critical US Cyber Protections

The social network's new privacy feature is technically flawed, opt-in, and limited in its functionality. All this for just $8 a month.

Elon Musk's long-promised launch of encrypted direct messages on Twitter has arrived. Like most attempts to add end-to-end encryption to a massive existing platform—never an easy proposition—there's good, bad, and ugly. The good: Twitter has added an optional layer of security for a small subset of its users that has never existed in Twitter's 16-plus years online. As for the bad and ugly: Well, that list is quite a lot longer.

On Wednesday night, Twitter announced the release of encrypted direct messages, a feature that Musk had assured users was coming from his very first days running the company. To Twitter's credit, it accompanied the new feature with an article on its help center breaking down the new feature's strengths and weaknesses with unusual transparency. And as the article points out, there are plenty of weaknesses. 

In fact, the company appears to have stopped short of calling the feature "end-to-end" encrypted, the term that would mean only users on the two ends of conversations can read messages, rather than hackers, government agencies that can eavesdrop on those messages, or even Twitter itself.

"As Elon Musk said, when it comes to Direct Messages, the standard should be, if someone puts a gun to our heads, we still can’t access your messages," the help desk page reads. "We’re not quite there yet, but we’re working on it."

In fact, the description of Twitter's encrypted messaging feature that follows that initial caveat seems almost like a laundry list of the most serious flaws in every existing end-to-end encrypted messaging app, now all combined into one product—along with a few extra flaws that are all its own.

The encryption feature is opt-in, for instance, not turned on by default, a decision for which Facebook Messenger has received criticism. It explicitly doesn't prevent “man-in-the-middle” attacks that would allow Twitter to invisibly spoof users' identities and intercept messages, long considered the most serious flaw in Apple's iMessage encryption. It doesn't have the “perfect forward secrecy” feature that makes spying on users harder even after a device is temporarily compromised. It doesn't allow for group messaging or even sending photos or videos. And perhaps most seriously, it currently restricts this subpar encrypted messaging system to only the verified users messaging each other—most of whom must pay $8 a month—vastly limiting the network that might use it.

“This clearly is not better than Signal or WhatsApp or anything that uses the Signal Protocol, in terms of features, in terms of security,” says Matthew Green, a professor of computer science at Johns Hopkins who focuses on cryptography, referring to the Signal Messenger app that's widely considered the modern standard in end-to-end encrypted calling and texting. Signal's encryption protocol is also used in both WhatsApp's encrypted-by-default communications and Facebook Messenger's opt-in encryption feature known as Secret Conversations. (Both Signal and WhatsApp are free, compared to the $8 per month for a Twitter Blue subscription that includes verification.) “You _should use those things instead_ if you really care about security,” Green says. "And they’ll be easier because you won’t have to pay $8 a month."

"On the positive side," Green adds, "hey, it’s a first step, maybe it’ll get better."

Musk has praised Signal in comments to Twitter's staff, and even said that he'd spoken with Signal's creator, Moxie Marlinspike, about similarly encrypting Twitter's DMs—a goal that Marlinspike himself shared when he briefly led Twitter's security team nearly a decade ago.

So Green—who has consulted at both WhatsApp and Facebook in their rollouts of encryption features based on Signal's protocol—was surprised to see that Twitter's encrypted messaging feature lacks so many of the positive properties of Signal and WhatsApp's end-to-end encryption. Beyond its lack of support for encrypted photos, videos, and group chats—key features of both Signal and WhatsApp—it also excludes the Signal protocol's constantly changing cryptographic keys, which are used to encrypt each message and never repeat.

That feature of Signal is what ensures "perfect forward secrecy," the security property that if a device is somehow compromised and the private key that decrypts messages is stolen, an eavesdropper still can't spy on future messages to and from that user. “I'm a little baffled by the lack of perfect forward secrecy,” says Green. “That's a basic feature of the Signal protocol.”

Twitter writes in its help center explanation that it essentially couldn't make that feature work while preserving the ability to access DMs when the user logs in on a new device. “We don’t plan to address this limitation,” the article reads.

Then there's the company's professed inability to stop “man-in-the-middle” attacks, in which Twitter itself could spoof users' identities to intercept their messages. In end-to-end encryption systems, messages are encrypted with an intended recipient's public key, such that only the recipient's private key—which is safely stored on the recipient's device—can decrypt them. But Twitter could trick a user—or even be compelled to do so by a government—so that their device invisibly encrypts messages to an eavesdropper's public key instead. Those messages could then be read and then re-encrypted with the intended recipient's key before they're sent on.

Apple's iMessage, which is otherwise considered a relatively strong end-to-end encryption system, has long suffered from this same vulnerability. But WhatsApp and Signal attempt to prevent man-in-the-middle attacks by allowing users to check a key “fingerprint” that ensures they're encrypting messages to the intended recipient. For now, Twitter has no such fingerprint-checking feature, though it says that it will add it soon.

That missing feature may be part of why Twitter has so far declined to even claim that it offers true end-to-end encryption, the "can't-read-your-messages-with-a-gun-to-our-head" feature Musk has promised.

"This appears to be a hasty deployment of a product that isn’t quite fully baked yet," says Riana Pfefferkorn, a security researcher at Stanford University's Internet Observatory. She points out that Zoom was penalized by the Federal Trade Commission in 2020 for claiming that it offered "end-to-end" encryption when it didn't—and that Twitter's reluctance to use the term may be a sign that it's not sure its system could meet that “end-to-end-encrypted” standard.

While Twitter is remarkably transparent about its encrypted DM feature's shortcomings on its help center page, Pfefferkorn worries that its flaws may not be as clear in the actual web and app interface that users see. “I think it was a good choice for the help page to try from the very first paragraph to manage expectations,” she says. “It remains to be seen whether Twitter users will believe that encrypted DMs offer more privacy and security than they actually do.”

Perhaps the most serious drawback to Twitter's encrypted DMs is simply that very few of its users will have the ability to send or receive them. The feature, at least for now, only works between two verified accounts, which either must be verified institutions or users who pay $8 a month for their blue check mark. “This shouldn’t be something you have to pay for,” says Green. “You shouldn’t have to pay for basic security.”

The notion of end-to-end encrypted Twitter DMs might one day offer a crucial new method to find someone online and send them a secret message; after all, Signal and WhatsApp's biggest drawback is that both require that you know a person's cell phone number, while Twitter DMs allow strangers to interact more freely. But as long as the encrypted DM feature is only available to send messages to and from verified accounts, its network will, by some measures, be even more restricted, limited to only a tiny fraction of Twitter's overall users.

For Twitter's security-conscious users, there remains only one way to send someone an encrypted message, and it hasn't changed in years. Send someone a DM, ask for their Signal number, and use Signal to start an actual, end-to-end encrypted conversation.

_Additional reporting by Lily Hay Newman_

Twitter’s Encrypted DMs Are Deeply Inferior to Signal and WhatsApp

We talk to the Signal Foundation’s Meredith Whittaker about how the surveillance economy is newer than we all might realize—and what we can do to fight back.

How to Reclaim Your Online Privacy

The unidentified attackers have targeted people on both sides of Russia’s war against Ukraine, carrying out espionage operations that suggest state funding.

Ukrainian networks have been on the receiving end of grimly sophisticated and innovative cyberattacks from pRussia for nearly a decade, and Ukraine has increasingly struck back, particularly since the Kremlin's invasion last year. Amidst all of this and activity from other governments and hacktivists, researchers from the security firm Malwarebytes say that they've been tracking a new hacking group that has been conducting espionage operations since 2020 against both pro-Ukraine targets in central Ukraine and pro-Russia targets in eastern Ukraine. 

Malwarebytes attributes five operations between 2020 and the present to the group, which it has dubbed Red Stinger, though the researchers only have insights into two of the campaigns conducted in the past year. The group's motives and allegiance aren't yet clear, but the digital campaigns are noteworthy for their persistence, aggressiveness, and lack of ties to other known actors.

The campaign that Malwarebytes calls “Operation Four” targeted a member of Ukraine's military who works on Ukrainian critical infrastructure, as well as other individuals whose potential intelligence value is less obvious. During this campaign, attackers compromised victims' devices to exfiltrate screenshots and documents, and even record audio from their microphones. In Operation Five, the group targeted multiple election officials running Russian referendums in disputed cities in Ukraine, including Donetsk and Mariupol. One target was an adviser to Russia's Central Election Commission, and another works on transportation—possibly railroad infrastructure—in the region.

“We were surprised about how big these targeted operations were, and they were able to gather a lot of information,” says Roberto Santos, a threat intelligence researcher at Malwarebytes. Santos collaborated on the investigation with former colleague Hossein Jazi, who first identified Red Stinger activity. “We have seen past targeted surveillance, but the fact that they were collecting real microphone recordings from victims and data from USB drives, it's unusual to see.”

Researchers from the security firm Kaspersky first published about Operation 5 in late March, naming the group behind it Bad Magic. Kaspersky similarly saw the group focusing on government and transportation targets in eastern Ukraine, along with agricultural targets.

“The malware and techniques used in this campaign are not particularly sophisticated, but are effective, and the code has no direct relation to any known campaigns,” Kaspersky researchers wrote.

The campaigns begin with phishing attacks to distribute malicious links that lead to tainted ZIP files, malicious documents, and special Windows linking files. From there, attackers deploy basic scripts to act as a backdoor and a loader for malware. The Malwarebytes researchers note that Red Stinger seems to have developed its own hacking tools and reuses characteristic scripts and infrastructure, including specific malicious URL generators and IP addresses. The researchers were able to expand their understanding of the group's operations after discovering two victims who appear to have infected themselves with Red Stinger malware while testing it.

“It's happened in the past with different attackers that they infect themselves,” Santos says. “I think they just got lazy because they were undetected since 2020.”

Red Stinger appears to be currently active. With details about its operations now entering the public sphere, the group may tweak its methods and tools in an attempt to evade detection. The Malwarebytes researchers say that by releasing information about the group's activities, they hope other organizations will deploy detections for Red Stinger operations and search their own telemetry for additional indications of what the hackers have done in the past and who is behind the group.

A Mysterious New Hacker Group, Red Stinger, Is Lurking in Ukraine’s Cyberspace

For a decade, a group called Big Pipes has worked behind the scenes with the FBI to target the worst cybercriminal “booter” services plaguing the internet.

When the FBI announced the takedown of 13 cyberattack-for-hire services yesterday, it may have seemed like just another day in law enforcement’s cat-and-mouse game with a criminal industry that has long plagued the internet’s infrastructure, bombarding victims with relentless waves of junk internet traffic to knock them offline. In fact, it was the latest win for a discreet group of detectives that has quietly worked behind the scenes for nearly a decade with the goal of ending that plague for good.

Yesterday’s operation was just the most recent of three major cybercriminal takedowns in the past five years that all began inside an informal working group that calls itself Big Pipes. The team’s roughly 30 members, who communicate mostly through Slack and weekly video calls, include staffers from several of the internet’s biggest cloud service providers and online gaming companies—though members from those companies spoke to WIRED on the condition that their employers not be named—as well as security researchers, academics, and a small number of FBI agents and federal prosecutors.

Big Pipes’ detectives have for years methodically tracked, measured, and ranked the output of “booter” or “stresser” services that sell distributed denial-of-service (DDOS) attacks that allow their customers to barrage enemies’ servers with disruptive floods of data. They’ve hunted the operators of those services, with private-sector members of the group often digging up leads that they hand to the group’s law enforcement agents and prosecutors. Together, they worked to initiate a takedown operation in December 2018 that led to the arrest of three hackers and knocked a dozen booter services offline. Last December, their work laid the foundation for Operation Power Off, which led to six arrests and the takedown of no fewer than 49 DDOS-for-hire sites, the biggest bust of its kind.

Yesterday’s takedowns, just four months after Operation Power Off, suggest the operations resulting from the group’s work may be accelerating. And Big Pipes is still tracking and hunting the booters that remain online, warns Richard Clayton, who leads a security research team at Cambridge University and has served as one of the group’s longest-running members. “We’re hoping that some of the people who were not taken down in this round get the message that perhaps it’s time they retired,” says Clayton. “If you weren’t seized this time, you might conclude you’ve pushed up your chance of being investigated. You might not want to wait and see what happens.”

Big Pipes Start Fights

The idea for Big Pipes was sparked at the Slam Spam conference in Pittsburgh in 2014, when Allison Nixon, a security researcher then at Deloitte, met with Elliot Peterson, an FBI agent who’d recently worked on the takedown of the notorious Game Over Zeus botnet. Nixon suggested to Peterson that they collaborate to take on the growing problem of booter services: At the time—and still today—hackers were wreaking havoc by launching ever-growing DDOS attacks across the internet for nihilistic fun, petty revenge, and profit, increasingly selling their attacks as a service.

In some cases, attackers would use botnets of thousands of computers infected with malware. In others, they’d use “reflection” or “amplification” attacks, exploiting servers run by legitimate online services that could be tricked into sending large amounts of traffic to an IP address of the hackers’ choosing. In many instances, gamers would pay a fee to one of a growing number of booter services—often just around $20 dollars for a subscription offering multiple attacks—to hit their rivals’ home connections. Those DDOS techniques frequently caused serious collateral damage for the internet service providers dealing with those indiscriminate floods of traffic. In some cases, DDOS attacks aimed at a single target could take down entire neighborhoods’ internet connections; disrupt emergency services; or, in one particularly gruesome case, break automated systems at a chicken farm, killing thousands of birds.

Big Pipes soon began to recruit staff from major internet services who had firsthand knowledge of booters based on their experiences as both victims and defenders in their attacks. (The group got its name from the phrase “big pipes start fights,” a joke about its members bragging about who among them had the biggest bandwidth on the internet.) Nixon and Clayton, for their part, contributed data from sensor networks they’d created—honeypots designed to join hackers’ botnets or act as their reflection servers and thus allow the researchers to see what attack commands the hackers were sending.

From Big Pipes’ inception, some members also went so far as to actively hunt for the identities of booter service operators, using clues from their forum posts and the websites where they advertised their attack services as starting points to try to unmask them. In one instance, a member of the group identified a booter operator by following a trail of online pseudonyms, phone numbers, and email addresses that led him from the hacker’s handle on the website HackForums—“itsfluffy”—to a web page that revealed his day job as a trainer for Pawfect Dog Training, along with his real name, Matthew Gatrel. “The operators of commodity DDOS services are not the most sophisticated actors out there,” says the Big Pipes member who followed those breadcrumbs, and who asked to remain unnamed. “They make mistakes.”

A Christmas Takedown Tradition

As Big Pipes’ data collection on booter service operators grew, so did the group’s partnership with the FBI. Eventually, that collaboration developed into an intermittent Christmas tradition of rounding up and disrupting as many of the internet’s worst booter services as possible. The timing of these operations, Big Pipes’ members emphasize, wasn’t intended for cruelty but as a response to the hackers’ own targeting of the holiday: For years, nihilistic hackers would wait until Christmas Day to launch disruptive DDOS attacks against online gaming services like the Playstation Network and Xbox Live, aiming to knock major gaming services offline on the busiest day of the year, just as kids were trying out their newly gifted games.

So in 2018, Big Pipes’ members worked with the FBI and the US Justice Department to stage their own pre-Christmas intervention, sifting through their data and giving leads to the group’s agents and prosecutors to take out the most active services in the growing booter industry. “We’re figuring out target selection: Which of these booter owners can be identified? Which of these booters are the highest harm in terms of the amount of DDOS traffic they’re pushing?” says Nixon, who today works at the security firm Unit221b. “So we figure out, OK, these are the highest-harm targets, these ones are low-hanging fruit. Who are we actually going to take down?”

In December of 2018, just five days before Christmas, the FBI announced a bust of 15 of the booters Big Pipes had suggested were the worst offenders. They included one called Quantum that the FBI says had launched 80,000 DDOS attacks and another, DownThem, accused of launching no fewer than 200,000. Three men operating those services in Pennsylvania, California, and Illinois—including the dog trainer Matthew Gatrel—were arrested and charged.

In the wake of that operation, Clayton’s Cambridge research team found that attacks from booter services fell by nearly a third for more than two months, and the services’ attacks with US victims were nearly cut in half for that time. So Big Pipes suggested they do it all again, only now going after _every_ major booter service that remained online. “Let’s see what happens if we go after everything that matters,” says Peterson, the FBI agent. “How do they react?”

It would take four years for the FBI and Justice Department to work back up to a second major booter takedown, following long delays that included Gatrel’s trial—he was sentenced in 2021 to two years in prison—and the Covid-19 pandemic. But finally, last December the FBI pulled off an even bigger purge of the booter underworld. Along with UK and Dutch federal police, they arrested six booter operators and tore down 49 web domains for booter services—all based on a long list of targets assembled from Big Pipes’ data about the most prominent and high-volume cyberattack services.

In fact, Clayton says that the operation took offline 17 of the top 20 booter services, based on his Cambridge research team’s data. Among the larger list of targets of the operation, he found that half of the 49 services returned under new names, but they carried out only half as much attack traffic for the next several months, with the number of attacks only returning to their previous level in March. That sustained dip was due, Clayton guesses, to the deterrent effect of the operation on potential booter customers. “I’d been pushing this idea that we should take down every booter in the world,” Clayton says. “We got halfway there.”

Yesterday, the FBI and Justice Department announced the success of yet another mass booter takedown, this time seizing 13 web domains of booter services. In fact, the DOJ says that 10 of those domains were seizures of reincarnated, renamed booters that had also been seized in the previous sweep in December, an action meant to signal to booter operators that they can’t evade law enforcement by merely relaunching their service with a new name and domain. Meanwhile, prosecutors also announced yesterday that four of the six defendants charged in that previous operation have now pleaded guilty.

Honeypots, Google Ads, Knock-and-Talks

Despite their constant communication, the members of Big Pipes and the FBI are careful to note that the internet services with staff members in the group don’t share their users’ private information without going through the usual legal processes of subpoenas and search warrants. Nor does the FBI share private data with Big Pipes, or blindly arrest or search people based on the group’s leads, Peterson says; the FBI investigates the defendants from scratch, treating information from Big Pipes as it would tips from any source. The FBI’s 2018 case against Gatrel, for instance, began with a subpoena to Cloudflare—a DDOS mitigation service Gatrel was ironically using to protect his own booter website—and then search warrants for Gatrel’s Google accounts.

But Peterson says Big Pipes’ work has nonetheless significantly helped him understand who to target in the booter landscape and how to pursue them far more efficiently. “If you take Big Pipes away, could we have worked cases against booter services? Yes,” he says. “But it might have taken a few more years to get to a similar scale.”

The FBI and Big Pipes’ increasing tempo of disruption may well just push booter services deeper into the shadows, rather than eliminating them. But if booter operators stop advertising on the open internet and move to the dark web, for instance, Clayton argues that the move would make it more obvious to their customers that the services are illegal and risky, and thus reduce demand for them.

In fact, he and other members of Big Pipes argue that most booter customers seem to believe—or convince themselves—that merely paying to use one of the services to knock out an adversary’s internet connection isn’t against the law, or at least isn’t an enforceable crime. When the UK’s National Crime Agency (NCA) ran a six-month Google advertising campaign in 2018 to intercept people seeking booter services and warn them about their illegality, Clayton’s research group found that attack traffic in the UK remained flat for those six months, while it increased at its usual pace in other countries.

Courtesy of FBI

In the years since, law enforcement agencies seem to have learned from that experiment: The FBI now also buys similar Google advertisements to warn potential booter customers that paying for the services is a crime. The UK’s NCA, meanwhile, has not only launched new advertising campaigns but even run its own fake booter services to identify would-be customers and then send them warnings—sometimes even with in-person visits—about the consequences of paying for criminal DDOS attacks.

Big Pipes’ Allison Nixon says she hopes that softer tactics like those can intercept would-be booter service operators early, before they start committing felonies: She’s found that most booter operators start as customers before launching their own service. But for people who aren’t dissuaded by those interventions, she says, Big Pipes and its partners at the FBI will still be watching them.

“The hope is that this whole show of force will convince some of them to quit and get a real job,” Nixon says. “We want to send a message that there are people tracking you. There are people paying attention to you. We have our eyes on you, we might get you next. And it might not even be on Christmas.”

The Team of Sleuths Quietly Hunting Cyberattack-for-Hire Services

OpenAI has new tools that give you more control over your information—although they may not go far enough.

There’s a chance that ChatGPT knows personal details about you—and if it doesn’t, it might just make something up. As OpenAI’s generative text chatbot has boomed in popularity over the past six months, the risks of the system being trained on data vacuumed up from the web have become clearer.

Data regulators around the world are investigating issues with how OpenAI gathered the data it uses to train its large language models, the accuracy of answers it provides about people, and other legal concerns about the use of its generative text systems. Europe’s data regulators have joined forces to look at OpenAI after Italy temporarily banned ChatGPT from the country. And Canada is also investigating the technology’s potential privacy risks.

In Europe, GDPR laws require companies and organizations to demonstrate lawful reasons for handling people’s personal information and to let people access information about them, be informed of how their information is used, and demand that errors be rectified. In some cases, they can ask that certain types of data be erased. The way people’s personal information has been used in training data has been an early area of concern for EU regulators.

As people have experimented with the chatbot, asking it questions about their lives and friends, a range of potential problems have emerged. OpenAI warns that ChatGPT may provide inaccurate information, and people have found that it makes up jobs and hobbies. It has cooked up false newspaper articles that had even the alleged human authors wondering if they were real. It generated incorrect statements saying a law professor was involved in a sexual harassment scandal, and it said a mayor in Australia had been implicated in a bribery scandal—he is preparing to sue for defamation.

It’s not just individuals who are concerned about how data is used. Samsung has banned employees from using generative AI tools, in part over fears about how data is stored on external servers and the risk that company secrets could ultimately be disclosed to other users. (There are separate issues around copyright and intellectual property.)

In response to the scrutiny—particularly from the Italian data regulator, which has now allowed ChatGPT back into the country after OpenAI made changes to its service—the company has introduced tools and processes that allow people more control over at least some of their data. Here’s how to use them.

Get Your Data Removed From ChatGPT

ChatGPT and GPT-4 generate their human-like answers statistically—predicting which words are likely to follow others after seeing millions of examples of sentences written by human authors. OpenAI has been secretive about the data it has trained its large language models on, so nobody outside the company knows exactly how much of the web (including people’s personal information) it has scraped in the process.

OpenAI says its large language models are trained on three sources of information: data taken from the web, data that the company licenses from others, and the information people feed it through chats. This can include information about individuals. “A large amount of data on the internet relates to people, so our training information does incidentally include personal information,” OpenAI explains in a post, stating that it takes steps to reduce the amount it gathers.

OpenAI has now introduced a Personal Data Removal Request form that allows people—primarily in Europe, although also in Japan—to ask that information about them be removed from OpenAI’s systems. It is described in an OpenAI blog post about how the company develops its language models.

The form primarily appears to be for requesting that information be removed from answers ChatGPT provides to users, rather than from its training data. It asks you to provide your name; email; the country you are in; whether you are making the application for yourself or on behalf of someone else (for instance a lawyer making a request for a client); and whether you are a public person, such as a celebrity.

OpenAI then asks for evidence that its systems have mentioned you. It asks you to provide “relevant prompts” that have resulted in you being mentioned and also for any screenshots where you are mentioned. “To be able to properly address your requests, we need clear evidence that the model has knowledge of the data subject conditioned on the prompts,” the form says. It asks you to swear that the details are correct and that you understand OpenAI may not, in all cases, delete the data. The company says it will balance “privacy and free expression” when making decisions about people’s deletion requests.

Daniel Leufer, a senior policy analyst at digital rights nonprofit Access Now, says the changes that OpenAI has made in recent weeks are OK but that it is only dealing with “the low-hanging fruit” when it comes to data protection. “They still have done nothing to address the more complex, systemic issue of how people’s data was used to train these models, and I expect that this is not an issue that’s just going to go away, especially with the creation of the EDPB taskforce on ChatGPT,” Leufer says, referring to the European regulators coming together to look at OpenAI.

“Individuals also may have the right to access, correct, restrict, delete, or transfer their personal information that may be included in our training information,” OpenAI’s help center page also says. To do this, it recommends emailing its data protection staff at dsar@openai.com. People who have already requested their data from OpenAI have not been impressed with its responses. And Italy’s data regulator says OpenAI claims it’s “technically impossible” to correct inaccuracies at the moment.

How to Delete Your ChatGPT Chat History

You should be cautious of what you tell ChatGPT, especially given OpenAI’s limited data-deletion options. The conversations you have with ChatGPT can, by default, be used by OpenAI in its future large language models as training data. This means the information could, at least theoretically, be reproduced in answer to people’s future questions. On April 25, the company introduced a new setting to allow anyone to stop this process, no matter where in the world they are.

When logged in to ChatGPT, click on your user profile in the bottom left-hand corner of the screen, click **Settings**, and then **Data Controls**. Here you can **toggle off Chat History & Training**. OpenAI says turning your chat history off means data you input into conversations “won’t be used to train and improve our models.”

As a result, anything you enter into ChatGPT—such as information about yourself, your life, and your work—shouldn’t be resurfaced in future iterations of OpenAI’s large language models. OpenAI says when chat history is turned off, it will retain all conversations for 30 days “to monitor for abuse” and then they will be permanently deleted.

When your data history is turned off, ChatGPT nudges you to turn it back on by placing a button in the sidebar that gives you the option to enable chat history again—a stark contrast to the “off” setting buried in the settings menu.

How To Delete Your Data From ChatGPT

Documents obtained by WIRED show SafeGraph, which sold location data related to Planned Parenthood visits, is now pursuing contracts with the US Air Force.

In its early years, SafeGraph sold direct access to individualized location traces tied to device IDs. SafeGraph has historically denied any links to law enforcement. “Contrary to some belief, we don’t have any law enforcement customers,” Hoffman, the founder, wrote in 2022. However, this kind of data, sold by other vendors, has reportedly been purchased by the FBI, Immigration and Customs Enforcement, and local police. 

In 2019, it spun off a subsidiary company called Veraset, which took over SafeGraph’s relationships with data providers, such as apps and other data brokers, and its business selling individualized data. The SafeGraph brand then shifted to selling aggregated and processed products based in part on Veraset’s raw data.

The AFWERX contract is the first publicly reported relationship between SafeGraph and the US military, but the company has a history of working with other government agencies. In 2018, it sold two years of raw data to the Illinois Department of Transportation. In the first months of the Covid-19 pandemic, it inked a $420,000 deal with the Centers for Disease Control. Meanwhile, Veraset gave raw, individualized data about millions of people to the Washington, DC, Department of Health and other agencies around the country. And in 2020 and 2021, Santa Clara County used SafeGraph data to monitor attendance at a local church as part of a broader effort to enforce Covid restrictions. Materials shared with the Air Force mention relationships with the US Department of Agriculture, the Federal Reserve Bank, and the Los Angeles County, New York City, and New Jersey governments. 

SafeGraph has also shown interest in the homeland security business. In 2022, it cosponsored the Institute for Defense and Government Advancement’s Homeland Security Week, a trade show and expo that connects security contractors with top federal and local law enforcement officials. In a joint video presentation with Department of Homeland Security contractor GoTenna, a SafeGraph representative advertised “strategic intelligence” to help border patrol agents monitor US “points of entry” and areas that can’t be watched in other ways. 

In May 2022, Vice revealed that SafeGraph was selling access to aggregated counts of where people were before and after visiting abortion clinics, including Planned Parenthood. In response, 14 US senators sent the company a letter demanding answers about its business practices, and SafeGraph promised to stop selling data about abortion clinic visitors.

Although SafeGraph is best known for dealing in cell-phone-based location data, its pitch to the Air Force makes little mention of data about human movement—the only direct reference is a slide that says it can help “analyze human activity for Landing Zone (LZ) selection,” without explaining what that means. But SafeGraph has recently expanded its business to incorporate other kinds of data as well. For example, in 2022, it launched Spend, a product that profiles the customers of brick-and-mortar stores, including what they spend, what wireless carriers they use, and whether they take out short-term “buy now, pay later” loans. 

In its slide deck for the Air Force, SafeGraph also claims to use a “global language model” able to “ingest data from 193 countries” and “100+ languages” to augment its data about physical places using “web-crawling and Machine Learning.” In response to a question about outstanding challenges, it said it plans to build “deeper language models” to help it gather data about places of interest to the Air Force, such as the Middle East.

SafeGraph is funded in part by the CIA-backed venture capital firm In-Q-Tel, and In-Q-Tel head of investments George Hoyem sits on SafeGraph’s board, according to its slide deck. SafeGraph has also received investments from a motley crew including Peter Thiel, Sam Harris, former Republican House majority leader Eric Cantor, and former Saudi Arabian intelligence chief Prince Turki bin Faisal Al Saud. SafeGraph’s last funding round valued the company at $370 million, according to the records.

In a post-contract questionnaire, SafeGraph told AFWERX it had met eight times with various parts of the Air Force to discuss the use of its products. It now plans to apply for a larger Phase 2 contract with the Air Force in the third quarter of 2023, targeting Task Force 99 of the Al Udeid Air Base in Qatar. The records also mention the possibility of a facility security clearance, which would allow SafeGraph to access classified information directly.

SafeGraph Lands US Air Force Contract After Targeting Abortion Clinics

A  few simple tools can help filter out most Twitter Blue users (but still see the ones you like).

I'm not here to judge anyone who pays for Twitter Blue. However, I'm not a fan of pay-to-win mechanics. At this point, Twitter is a game where players compete for the most attention; Twitter Blue is overpowered DLC. If you buy a subscription, your tweets are shown at the top of comment threads and prioritized in other contexts, including the “For You” page. This makes Blue the social media equivalent of paying for unlimited ammo or improved body armor, regardless of who you are or whether what you have to say is worth promoting.

It also makes Twitter really annoying to use. No one wants to play with the people who are paying to win. That's why there's so much mockery of, and desire to avoid, Twitter Blue users.

Granted, there are plenty of reasons why longtime Twitter power users—in particular, public figures or people who historically have trouble with impersonators—might pay for Twitter Blue. No one wants to lose an audience they worked hard to build, and Twitter has the right to monetize itself however it likes. You also have the right to consume (or not consume) the social media you choose.

All the same, it can't be denied that a lot of really annoying posts from Twitter Blue users with low follower counts show up in all kinds of contexts. If you don't want to see those posts, I don't blame you, and here's how you can filter them all out:

Blue Lite Blocker

Twitter via Justin Pot

BlueLiteBlocker is a free, open-source browser extension for Chrome and Firefox that can filter comments from Twitter Blue subscribers out of your Twitter timeline. You can install the extension and then use Twitter the way you normally do. All comments from Twitter Blue users will be obscured—you can click “View” if you are absolutely certain you want to see them.

Unlike some extensions (BluesBlocker, for example, or Twitter Blue Blocker), Blue Lite Blocker doesn't affect anyone that you actively follow, meaning it won't accidentally filter out your friends. The extension also allows you to still see the tweets from Twitter Blue users that have a certain number of followers—the default threshold is 100,000, but you can set it to whatever you want.

Twitter via Justin Pot

These two features make Blue Lite Blocker the perfect compromise. It filters out the annoying Twitter Blue users while not affecting the people you care about and the public figures whose posts you might want to see. If there's a Twitter Blue user with a lot of followers who you don't want to see, you can block or mute them manually.

Other Workarounds to Clean Up Twitter

Blue Lite Blocker isn't a perfect solution. It requires a browser extension, which not everyone is going to like, and it doesn’t help on mobile (not that there’s much help on mobile anyway). There are a few alternatives, though.

Twitter via Justin Pot

The first is to make sure you're using the Following tab instead of For You. This tab is just tweets from the people you follow, so it doesn't artificially favor Twitter Blue subscribers. You could also try blocking all retweets if you want to clean things up a little more.

If that's not enough, you could try TweetDeck, the power user's version of Twitter. This tool doesn't have a For You page or any algorithmic sorting, meaning Twitter Blue users aren't artificially boosted on it. TweetDeck is the one part of Twitter that's completely untouched by the recent changes there, and I hope this doesn't change. (So please, no one tell Elon Musk that TweetDeck still exists.)

Another cool extension, if you miss seeing legacy checkmarks, is Eight Dollars. Right now blue checkmarks only show up if someone is paying for Blue, meaning most people who had a checkmark prior to April 20, 2023 no longer have one. Eight Dollars can show you which users were verified before legacy verified checkmarks were removed, which is useful if you trusted the old system.

There's a cat and mouse element to all of this, of course. Twitter's current regime really, really wants to turn Blue checkmarks into sellable status symbols and any tool that diminishes the visibility of Blue subscribers cuts against that goal. They're going to try to find a way to break this, and all similar, tools. For now, though, they work—use them while you can. 

Or you could just abandon your Twitter account, set up a public archive of your tweets, and learn how to get started on Mastodon or Bluesky. Up to you.

Your Twitter Feed Sucks Now. These Free Add-Ons Can Help

Plus: Apple and Google plan to stop AirTag stalking, Meta violated the FTC’s privacy order, and how to tell if your car is tracking you.

In December 2020, security giant Mandiant revealed it had been hacked. Its disclosure was the first public sign of the SolarWinds hack, a Russian-orchestrated supply chain attack that's widely regarded as one of the biggest espionage hacks ever. Among its victims were the US Departments of Homeland Security, Energy, and Justice. This blow-by-blow retelling of the historic SolarWinds attack, from Kim Zetter, charts the ways the hackers pulled off the attack—and how they were eventually caught.

Anti-abortion group the American College of Pediatricians (ACPeds) suffered a significant data breach this week. The doctors' organization, which sued the US government to ban the abortion drug mifepristone, left an unsecured Google Drive on its website, exposing a decade's worth of email exchanges, financial and tax records, and more sensitive data. The details give an unprecedented view of the organization, which has been described as a “hate group” for its views on LGBTQ people.  While ACPeds—which is not a school at all—characterizes itself as a “scientific organization,” leaked records show its deeply evangelical Christian mission.

Security experts have promised a future where passwords will cease to exist for the best part of a decade. However, that reality took a big step forward this week—really!—as Google launched passkey logins for billions of people. The technique uses cryptographic keys that are stored on your devices to replace your old, insecure passwords.

Elsewhere, cops in the US, Europe, and nine other countries have arrested 288 people for their involvement in the dark web drug markets, including the site Monopoly Market, which was quietly taken offline in 2021. Facebook owner Meta has added new tools to its business accounts in an attempt to thwart bad actors abusing them, including who can become account administrators and access lines of credit.

But that’s not all. Each week, we round up the news we didn’t report in-depth ourselves. Click on the headlines to read the full stories. And stay safe out there.

Russian ships with underwater operations equipment have been identified as being near the sites of the Nord Stream gas pipeline explosions in the days before the blasts, according to a joint investigation from national broadcasters in Denmark, Norway, Sweden, and Finland. Journalists at the publications combined intercepted radio broadcasts from the ships with satellite images to pinpoint their locations and track their paths. It is the latest example of investigators piecing together different sources of data, from varying unconnected sources, to reveal new details about real-world events.

Three ships, according to the investigation, sailed from naval bases in Russia to near the blast sites in June and September 2022. All of the ships had turned off their location tracking AIS services, an act often described as “going dark” and commonly used for disguising activity. Among the vessels were the navy research ship Sibiryakov and a tugboat called SB-123, which is said to be capable of launching mini-submarines. (In November 2022, WIRED reported on the presence of “ghost ships” around the time of the explosions, but had no information on their identity.)

Separately, another Russian vessel, the SS-750, was near the pipelines four days before they were blown up. In response to a public records request, the Danish Defense Command confirmed to the Information, a Danish news site, that it had 26 photos of the SS-750 near the sites.

Since the explosions at the Nord Stream 1 and 2 pipelines in September, there has been no official confirmation, with supporting evidence, of who may have been behind the blasts. The investigation from the Nordic journalists says the ships’ behavior was unusual but does not conclude what they were doing near the Nord Stream sites. Russia has denied being involved in the attacks, pointing fingers at both the UK and US. Other reports have claimed a pro-Ukranian group may have conducted the attack, which Ukraine has denied. Official investigations into the blasts are still ongoing in multiple European countries and the exact cause of the explosions is still unclear.

The US Federal Trade Commission is planning to impose a “blanket ban” on Meta and its companies—Instagram, Messenger, Facebook, WhatsApp, and VR firm Horizon Worlds—from making money from the data of people under 18. This could potentially mean the company would not be able to use data from these users to show them targeted advertisements and could only use the information it collects for providing its services or security purposes. The FTC also says that Meta would not be able to start using data collected before people turned 18 for commercial purposes.

The proposal comes as the FTC alleged Meta hasn’t complied with a previous 2020 privacy order it agreed to as part of a $5 billion settlement. In a statement, the FTC says Meta has “misled parents about their ability to control with whom their children communicated through its Messenger Kids app, and misrepresented the access it provided some app developers to private user data.” The proposed changes would be a modification to the FTC’s previous order, and it says it may also limit how Meta uses face recognition technology and require it to provide extra protection for users. Mark Zuckerberg’s firm has 30 days to respond to the FTC, which then may alter its plans.

How do you know if you are being stalked with an AirTag? If you have an iPhone running a newish version of iOS, then you should receive push notifications if an AirTag that doesn’t belong to you is following you around. However, if you own an Android device, you have to go through the extra step of downloading an app to discover if you are being tracked—a high burden for people in potentially abusive situations. Apple and Google said this week they are now working together to create a common standard for all phones to detect Bluetooth trackers that are being used to stalk or harass people. The proposal is also backed by Samsung and Bluetooth tracker manufacturers. The plan comes years after Apple’s cheap devices were first released without the additional privacy protections Apple added after the AirTags launch.

In 2017, the NotPetya cyberattack—the most devastating in history—caused more than $10 billion in damage to the companies and organizations it impacted. Among them was pharmaceutical giant Merck, which lost hundreds of millions of dollars as a result of the malware. This week, a New Jersey court ruled that Merck's insurers have to cover losses from the cyberattack. The insurers had argued the use of the malware, which has been linked to Russia, should be considered a warlike act and, as a result, would not fall under insurance policies.

The judges didn't agree. “The exclusion of damages caused by hostile or warlike action by a government or sovereign power in times of war or peace requires the involvement of military action,” they wrote in a decision. “Coverage could only be excluded here if we stretched the meaning of 'hostile' to its outer limit.” The decision comes as US president Joe Biden's administration looks to shift the liability for security issues to companies that are impacted.

Whether you know it or not, your car is a data goldmine. Within a few minutes, it is possible to "fingerprint" you based on the data your vehicle collects. Some consider the amount of data that vehicles grab a national security threat. A new free tool, from vehicle privacy company Privacy4Cars, aims to reveal how much data your car and its manufacturer can collect about you. Entering your vehicle identification number (VIN) into the tool will tell you whether your car collects "identifiers, location data, biometrics, and data synced from mobile phones," as well as what data manufacturers sell to third parties, such as data brokers.

Russian ‘Ghost Ships’ Identified Near the Nord Stream Blasts

Leaked documents reveal that the American College of Pediatricians viewed “mainstream medicine” and “nominal Christians” as its opposition.

In a declaration in federal court in 2021, Quentin Van Meter, the president of the American College of Pediatricians, described the organization as a “secular, scientific medical association,” whose “views are not religious as such.” 

ACPeds has also targeted donors, physicians, and other clinicians based solely on their political leanings, documents show. The group maintains a list of more than 5,000 “conservative doctors,” for instance, and records reveal that they have been routinely targeted with mailers designed to elicit interest in membership. 

In 2021, ACPeds solicited a proposal from a direct-mail fundraising agency on how to boost its fundraising efforts and where to spend the windfall. The agency recommended that ACPeds target “30,000 prospective conservative donors,” whose gifts, it said, would in turn be spent to “target conservative professionals in the medical community.” In an explanation of the services it offers, the agency said it could procure donor lists from “other like-minded organizations,” and that, if ACPeds wished, it could facilitate “exchanges” and “rentals” of other mailing lists.

Information about medical professionals would not be sold or exchanged, the agency said, if the professionals are “existing ACPeds members.” A contract between the group and the agency was finalized in August 2021, records show. 

Despite its homages to science, the views of ACPeds and its board are deeply rooted in a morality based exclusively in evangelical religious beliefs. Notes taken at board meetings, which open and close with prayer, show that its directors view consensus science, people who hold advanced degrees, and even the law itself as a threat to its agenda. Prayer is prescribed as “armor” against the group’s perceived adversaries, which include other Christians whose devotion they’ve judged to be inadequate.

Minutes from a 2017 board meeting read: “Threats to the College include the Southern Poverty Law Center (SPLC), the Human Rights Campaign (HRC), and the LGBTQ lobbying body, as well as mainstream medicine, psychology, academia, media, corporate America and nominal Christians, churches and organizations.” 

The atmosphere at ACPeds’ closed-door meetings, dozens of which are meticulously documented, contrasts sharply with the image it aims to project publicly. Conversations over exactly how religious the group can present itself publicly have rolled from one year to the next. During meetings in 2014 and 2015, members discussed the potential benefits of declaring their “recognition of God versus standing purely as a scientific organization.” A minute taker noted that no “definitive agreement” could be reached about “whether or not to do this.” 

Records of its membership totals show that as few as half of ACPeds’ 700 members may be practicing pediatricians, with its numbers being boosted by subscriptions from students, retirees, and so-called “friends” of the organization. Records show the group has also explored expanding its ranks to include additional members without medical expertise in response to its lackluster returns on expensive recruitment efforts. 

Debates over whether to harness its religiosity in a more public way have been attended by a Catholic attorney, who in 2014 advised them to “express belief in a deity without being evangelical.” After the opening prayer at the following year’s gathering, the group’s then president, Michelle Cretella, mimicked the lawyer’s advice, reminding members that ACPeds is not a “religious organization,” but a “theistic” one that recognizes “natural laws” imparted by a supreme being.

Source

Wired