Detecting Cloned Cards at the ATM, Register

May 14, 2018

Much of the fraud involving counterfeit credit, ATM debit and retail gift cards relies on the ability of thieves to use cheap, widely available hardware to encode stolen data onto any card’s magnetic stripe. But new research suggests retailers and ATM operators could reliably detect counterfeit cards using a simple technology that flags cards which appear to have been altered by such tools.

A gift card purchased at retail with an unmasked PIN hidden behind a paper sleeve. Such PINs can be easily copied by an adversary, who waits until the card is purchased to steal the card’s funds. Image: University of Florida.

Researchers at the University of Florida found that account data encoded on legitimate cards is invariably written using quality-controlled, automated facilities that tend to imprint the information in uniform, consistent patterns.

Cloned cards, however, usually are created by hand with inexpensive encoding machines, and as a result feature far more variance or “jitter” in the placement of digital bits on the card’s stripe.

Gift cards can be extremely profitable and brand-building for retailers, but gift card fraud creates a very negative shopping experience for consumers and a costly conundrum for retailers. The FBI estimates that while gift card fraud makes up a small percentage of overall gift card sales and use, approximately $130 billion worth of gift cards are sold each year.

One of the most common forms of gift card fraud involves thieves tampering with cards inside the retailer’s store — before the cards are purchased by legitimate customers. Using a handheld card reader, crooks will swipe the stripe to record the card’s serial number and other data needed to duplicate the card.

If there is a PIN on the gift card packaging, the thieves record that as well. In many cases, the PIN is obscured by a scratch-off decal, but gift card thieves can easily scratch those off and then replace the material with identical or similar decals that are sold very cheaply by the roll online.

“They can buy big rolls of that online for almost nothing,” said Patrick Traynor, an associate professor of computer science at the University of Florida. “Retailers we’ve worked with have told us they’ve gone to their gift card racks and found tons of this scratch-off stuff on the ground near the racks.”

At this point the cards are still worthless because they haven’t yet been activated. But armed with the card’s serial number and PIN, thieves can simply monitor the gift card account at the retailer’s online portal and wait until the cards are paid for and activated at the checkout register by an unwitting shopper.

Once a card is activated, thieves can encode that card’s data onto any card with a magnetic stripe and use that counterfeit to purchase merchandise at the retailer. The stolen goods typically are then sold online or on the street. Meanwhile, the person who bought the card (or the person who received it as a gift) finds the card is drained of funds when they eventually get around to using it at a retail store.

The top two gift cards show signs that someone previously peeled back the protective sticker covering the redemption code. Image: Flint Gatrell.

Traynor and a team of five other University of Florida researchers partnered with retail giant WalMart to test their technology, which Traynor said can be easily and quite cheaply incorporated into point-of-sale systems at retail store cash registers. They said the WalMart trial demonstrated that researchers’ technology distinguished legitimate gift cards from clones with up to 99.3 percent accuracy.

While impressive, that rate still means the technology could still generate a “false positive” — erroneously flagging a legitimate customer as using a fraudulently obtained gift card in a non-trivial number of cases. But Traynor said the retailers they spoke with in testing their equipment all indicated they would welcome any additional tools to curb the incidence of gift card fraud.

“We’ve talked with quite a few retail loss prevention folks,” he said. “Most said even if they can simply flag the transaction and make a note of the person [presenting the cloned card] that this would be a win for them. Often, putting someone on notice that loss prevention is watching is enough to make them stop — at least at that store. From our discussions with a few big-box retailers, this kind of fraud is probably their newest big concern, although they don’t talk much about it publicly. If the attacker does any better than simply cloning the card to a blank white card, they’re pretty much powerless to stop the attack, and that’s a pretty consistent story behind closed doors.” Continue reading

Think You’ve Got Your Credit Freezes Covered? Think Again.

May 9, 2018

I spent a few days last week speaking at and attending a conference on responding to identity theft. The forum was held in Florida, one of the major epicenters for identity fraud complaints in United States. One gripe I heard from several presenters was that identity thieves increasingly are finding ways to open new mobile phone accounts in the names of people who have already frozen their credit files with the big-three credit bureaus. Here’s a look at what may be going on, and how you can protect yourself.

Carrie Kerskie is director of the Identity Fraud Institute at Hodges University in Naples. A big part of her job is helping local residents respond to identity theft and fraud complaints. Kerskie said she’s had multiple victims in her area recently complain of having cell phone accounts opened in their names even though they had already frozen their credit files at the big three credit bureausEquifax, Experian and Trans Union (as well as distant fourth bureau Innovis).

The freeze process is designed so that a creditor should not be able to see your credit file unless you unfreeze the account. A credit freeze blocks potential creditors from being able to view or “pull” your credit file, making it far more difficult for identity thieves to apply for new lines of credit in your name.

But Kerskie’s investigation revealed that the mobile phone merchants weren’t asking any of the four credit bureaus mentioned above. Rather, the mobile providers were making credit queries with the National Consumer Telecommunications and Utilities Exchange (NCTUE), or nctue.com.

Source: nctue.com

“We’re finding that a lot of phone carriers — even some of the larger ones — are relying on NCTUE for credit checks,” Kerskie said. “It’s mainly phone carriers, but utilities, power, water, cable, any of those, they’re all starting to use this more.”

The NCTUE is a consumer reporting agency founded by AT&T in 1997 that maintains data such as payment and account history, reported by telecommunication, pay TV and utility service providers that are members of NCTUE.

Who are the NCTUE’s members? If you call the 800-number that NCTUE makes available to get a free copy of your NCTUE credit report, the option for “more information” about the organization says there are four “exchanges” that feed into the NCTUE’s system: the NCTUE itself; something called “Centralized Credit Check Systems“; the New York Data Exchange; and the California Utility Exchange.

According to a partner solutions page at Verizon, the New York Data Exchange is a not-for-profit entity created in 1996 that provides participating exchange carriers with access to local telecommunications service arrears (accounts that are unpaid) and final account information on residential end user accounts.

The NYDE is operated by Equifax Credit Information Services Inc. (yes, that Equifax). Verizon is one of many telecom providers that use the NYDE (and recall that AT&T was the founder of NCTUE).

The California Utility Exchange collects customer payment data from dozens of local utilities in the state, and also is operated by Equifax (Equifax Information Services LLC).

Google has virtually no useful information available about an entity called Centralized Credit Check Systems. It’s possible it no longer exists. If anyone finds differently, please leave a note in the comments section.

When I did some more digging on the NCTUE, I discovered…wait for it…Equifax also is the sole contractor that manages the NCTUE database. The entity’s site is also hosted out of Equifax’s servers. Equifax’s current contract to provide this service expires in 2020, according to a press release posted in 2015 by Equifax. Continue reading

Advertisement

Microsoft Patch Tuesday, May 2018 Edition

May 8, 2018

Microsoft today released a bundle of security updates to fix at least 67 holes in its various Windows operating systems and related software, including one dangerous flaw that Microsoft warns is actively being exploited. Meanwhile, as it usually does on Microsoft’s Patch Tuesday — the second Tuesday of each month — Adobe has a new Flash Player update that addresses a single but critical security weakness.

First, the Flash Tuesday update, which brings Flash Player to v. 29.0.0.171. Some (present company included) would argue that Flash Player is itself “a single but critical security weakness.” Nevertheless, Google Chrome and Internet Explorer/Edge ship with their own versions of Flash, which get updated automatically when new versions of these browsers are made available.

You can check if your browser has Flash installed/enabled and what version it’s at by pointing your browser at this link. Adobe is phasing out Flash entirely by 2020, but most of the major browsers already take steps to hobble Flash. And with good reason: It’s a major security liability. Continue reading

Study: Attack on KrebsOnSecurity Cost IoT Device Owners $323K

May 7, 2018

A monster distributed denial-of-service attack (DDoS) against KrebsOnSecurity.com in 2016 knocked this site offline for nearly four days. The attack was executed through a network of hacked “Internet of Things” (IoT) devices such as Internet routers, security cameras and digital video recorders. A new study that tries to measure the direct cost of that one attack for IoT device users whose machines were swept up in the assault found that it may have cost device owners a total of $323,973.75 in excess power and added bandwidth consumption.

My bad.

But really, none of it was my fault at all. It was mostly the fault of IoT makers for shipping cheap, poorly designed products (insecure by default), and the fault of customers who bought these IoT things and plugged them onto the Internet without changing the things’ factory settings (passwords at least.)

The botnet that hit my site in Sept. 2016 was powered by the first version of Mirai, a malware strain that wriggles into dozens of IoT devices left exposed to the Internet and running with factory-default settings and passwords. Systems infected with Mirai are forced to scan the Internet for other vulnerable IoT devices, but they’re just as often used to help launch punishing DDoS attacks.

By the time of the first Mirai attack on this site, the young masterminds behind Mirai had already enslaved more than 600,000 IoT devices for their DDoS armies. But according to an interview with one of the admitted and convicted co-authors of Mirai, the part of their botnet that pounded my site was a mere slice of firepower they’d sold for a few hundred bucks to a willing buyer. The attack army sold to this ne’er-do-well harnessed the power of just 24,000 Mirai-infected systems (mostly security cameras and DVRs, but some routers, too).

These 24,000 Mirai devices clobbered my site for several days with data blasts of up to 620 Gbps. The attack was so bad that my pro-bono DDoS protection provider at the time — Akamai — had to let me go because the data firehose pointed at my site was starting to cause real pain for their paying customers. Akamai later estimated that the cost of maintaining protection against my site in the face of that onslaught would have run into the millions of dollars.

We’re getting better at figuring out the financial costs of DDoS attacks to the victims (5, 6 or 7 -digit dollar losses) and to the perpetrators (zero to hundreds of dollars). According to a report released this year by DDoS mitigation giant NETSCOUT Arbor, fifty-six percent of organizations last year experienced a financial impact from DDoS attacks for between $10,000 and $100,000, almost double the proportion from 2016.

But what if there were also a way to work out the cost of these attacks to the users of the IoT devices which get snared by DDos botnets like Mirai? That’s what researchers at University of California, Berkeley School of Information sought to determine in their new paper, “rIoT: Quantifying Consumer Costs of Insecure Internet of Things Devices.

If we accept the UC Berkeley team’s assumptions about costs borne by hacked IoT device users (more on that in a bit), the total cost of added bandwidth and energy consumption from the botnet that hit my site came to $323,973.95. This may sound like a lot of money, but remember that broken down among 24,000 attacking drones the per-device cost comes to just $13.50.

So let’s review: The attacker who wanted to clobber my site paid a few hundred dollars to rent a tiny portion of a much bigger Mirai crime machine. That attack would likely have cost millions of dollars to mitigate. The consumers in possession of the IoT devices that did the attacking probably realized a few dollars in losses each, if that. Perhaps forever unmeasured are the many Web sites and Internet users whose connection speeds are often collateral damage in DDoS attacks.

Image: UC Berkeley.

Continue reading

Twitter to All Users: Change Your Password Now!

May 3, 2018

Twitter just asked all 300+ million users to reset their passwords, citing the exposure of user passwords via a bug that stored passwords in plain text — without protecting them with any sort of encryption technology that would mask a Twitter user’s true password. The social media giant says it has fixed the bug and that so far its investigation hasn’t turned up any signs of a breach or that anyone misused the information. But if you have a Twitter account, please change your account password now.

Or if you don’t trust links in blogs like this (I get it) go to Twitter.com and change it from there. And then come back and read the rest of this. We’ll wait.

In a post to its company blog this afternoon, Twitter CTO Parag Agrawal wrote:

“When you set a password for your Twitter account, we use technology that masks it so no one at the company can see it. We recently identified a bug that stored passwords unmasked in an internal log. We have fixed the bug, and our investigation shows no indication of breach or misuse by anyone.

A message posted this afternoon (and still present as a pop-up) warns all users to change their passwords.

“Out of an abundance of caution, we ask that you consider changing your password on all services where you’ve used this password. You can change your Twitter password anytime by going to the password settings page.” Continue reading

When Your Employees Post Passwords Online

May 2, 2018

Storing passwords in plaintext online is never a good idea, but it’s remarkable how many companies have employees who are doing just that using online collaboration tools like Trello.com. Last week, KrebsOnSecurity notified a host of companies that employees were using Trello to share passwords for sensitive internal resources. Among those put at risk by such activity included an insurance firm, a state government agency and ride-hailing service Uber.

By default, Trello boards for both enterprise and personal use are set to either private (requires a password to view the content) or team-visible only (approved members of the collaboration team can view).

But that doesn’t stop individual Trello users from manually sharing personal boards that include proprietary employer data, information that may be indexed by search engines and available to anyone with a Web browser. And unfortunately for organizations, far too many employees are posting sensitive internal passwords and other resources on their own personal Trello boards that are left open and exposed online.

A personal Trello board created by an Uber employee included passwords that might have exposed sensitive internal company operations.

KrebsOnSecurity spent the past week using Google to discover unprotected personal Trello boards that listed employer passwords and other sensitive data. Pictured above was a personal board set up by some Uber developers in the company’s Asia-Pacific region, which included passwords needed to view a host of internal Google Documents and images.

Uber spokesperson Melanie Ensign said the Trello board in question was made private shortly after being notified by this publication, among others. Ensign said Uber found the unauthorized Trello board exposed information related to two users in South America who have since been notified.

“We had a handful of members in random parts of the world who didn’t realize they were openly sharing this information,” Ensign said. “We’ve reached out to these teams to remind people that these things need to happen behind internal resources. Employee awareness is an ongoing challenge, We may have dodged a bullet here, and it definitely could have been worse.”

Ensign said the initial report about the exposed board came through the company’s bug bounty program, and that the person who reported it would receive at least the minimum bounty amount — $500 — for reporting the incident (Uber hasn’t yet decided whether the award should be higher for this incident).

The Uber employees who created the board “used their work email to open a public board that they weren’t supposed to,” Ensign said. “They didn’t go through our enterprise account to create that. We first found out about it through our bug bounty program, and while it’s not technically a vulnerability in our products, it’s certainly something that we would pay for anyway. In this case, we got multiple reports about the same thing, but we always pay the first report we get.”

Of course, not every company has a bug bounty program to incentivize the discovery and private reporting of internal resources that may be inadvertently exposed online.

Screenshots that KrebsOnSecurity took of many far more shocking examples of employees posting dozens of passwords for sensitive internal resources are not pictured here because the affected parties still have not responded to alerts provided by this author.
Continue reading

Security Trade-Offs in the New EU Privacy Law

April 27, 2018

On two occasions this past year I’ve published stories here warning about the prospect that new European privacy regulations could result in more spams and scams ending up in your inbox. This post explains in a question and answer format some of the reasoning that went into that prediction, and responds to many of the criticisms leveled against it.

Before we get to the Q&A, a bit of background is in order. On May 25, 2018 the General Data Protection Regulation (GDPR) takes effect. The law, enacted by the European Parliament, requires companies to get affirmative consent for any personal information they collect on people within the European Union. Organizations that violate the GDPR could face fines of up to four percent of global annual revenues.

In response, the Internet Corporation for Assigned Names and Numbers (ICANN) — the nonprofit entity that manages the global domain name system — has proposed redacting key bits of personal data from WHOIS, the system for querying databases that store the registered users of domain names and blocks of Internet address ranges (IP addresses).

Under current ICANN rules, domain name registrars should collect and display a variety of data points when someone performs a WHOIS lookup on a given domain, such as the registrant’s name, address, email address and phone number. Most registrars offer a privacy protection service that shields this information from public WHOIS lookups; some registrars charge a nominal fee for this service, while others offer it for free.

But in a bid to help registrars comply with the GDPR, ICANN is moving forward on a plan to remove critical data elements from all public WHOIS records. Under the new system, registrars would collect all the same data points about their customers, yet limit how much of that information is made available via public WHOIS lookups.

The data to be redacted includes the name of the person who registered the domain, as well as their phone number, physical address and email address. The new rules would apply to all domain name registrars globally.

ICANN has proposed creating an “accreditation system” that would vet access to personal data in WHOIS records for several groups, including journalists, security researchers, and law enforcement officials, as well as intellectual property rights holders who routinely use WHOIS records to combat piracy and trademark abuse.

But at an ICANN meeting in San Juan, Puerto Rico last month, ICANN representatives conceded that a proposal for how such a vetting system might work probably would not be ready until December 2018. Assuming ICANN meets that deadline, it could be many months after that before the hundreds of domain registrars around the world take steps to adopt the new measures.

In a series of posts on Twitter, I predicted that the WHOIS changes coming with GDPR will likely result in a noticeable increase in cybercrime — particularly in the form of phishing and other types of spam. In response to those tweets, several authors on Wednesday published an article for Georgia Tech’s Internet Governance Project titled, “WHOIS afraid of the dark? Truth or illusion, let’s know the difference when it comes to WHOIS.”

The following Q&A is intended to address many of the more misleading claims and assertions made in that article.

Cyber criminals don’t use their real information in WHOIS registrations, so what’s the big deal if the data currently available in WHOIS records is no longer in the public domain after May 25?

I can point to dozens of stories printed here — and probably hundreds elsewhere — that clearly demonstrate otherwise. Whether or not cyber crooks do provide their real information is beside the point. ANY information they provide — and especially information that they re-use across multiple domains and cybercrime campaigns — is invaluable to both grouping cybercriminal operations and in ultimately identifying who’s responsible for these activities.

To understand why data reuse in WHOIS records is so common among crooks, put yourself in the shoes of your average scammer or spammer — someone who has to register dozens or even hundreds or thousands of domains a week to ply their trade. Are you going to create hundreds or thousands of email addresses and fabricate as many personal details to make your WHOIS listings that much harder for researchers to track? The answer is that those who take this extraordinary step are by far and away the exception rather than the rule. Most simply reuse the same email address and phony address/phone/contact information across many domains as long as it remains profitable for them to do so.

This pattern of WHOIS data reuse doesn’t just extend across a few weeks or months. Very often, if a spammer, phisher or scammer can get away with re-using the same WHOIS details over many years without any deleterious effects to their operations, they will happily do so. Why they may do this is their own business, but nevertheless it makes WHOIS an incredibly powerful tool for tracking threat actors across multiple networks, registrars and Internet epochs.

All domain registrars offer free or a-la-carte privacy protection services that mask the personal information provided by the domain registrant. Most cybercriminals — unless they are dumb or lazy — are already taking advantage of these anyway, so it’s not clear why masking domain registration for everyone is going to change the status quo by much. 

It is true that some domain registrants do take advantage of WHOIS privacy services, but based on countless investigations I have conducted using WHOIS to uncover cybercrime businesses and operators, I’d wager that cybercrooks more often do not use these services. Not infrequently, when they do use WHOIS privacy options there are still gaps in coverage at some point in the domain’s history (such as when a registrant switches hosting providers) which are indexed by historic WHOIS records and that offer a brief window of visibility into the details behind the registration.

This is demonstrably true even for organized cybercrime groups and for nation state actors, and these are arguably some of the most sophisticated and savvy cybercriminals out there.

It’s worth adding that if so many cybercrooks seem nonchalant about adopting WHOIS privacy services it may well be because they reside in countries where the rule of law is not well-established, or their host country doesn’t particularly discourage their activities so long as they’re not violating the golden rule — namely, targeting people in their own backyard. And so they may not particularly care about covering their tracks. Or in other cases they do care, but nevertheless make mistakes or get sloppy at some point, as most cybercriminals do.

The GDPR does not apply to businesses — only to individuals — so there is no reason researchers or anyone else should be unable to find domain registration details for organizations and companies in the WHOIS database after May 25, right?

It is true that the European privacy regulations as they relate to WHOIS records do not apply to businesses registering domain names. However, the domain registrar industry — which operates on razor-thin profit margins and which has long sought to be free from any WHOIS requirements or accountability whatsoever — won’t exactly be tripping over themselves to add more complexity to their WHOIS efforts just to make a distinction between businesses and individuals.

As a result, registrars simply won’t make that distinction because there is no mandate that they must. They’ll just adopt the same WHOIS data collection and display polices across the board, regardless of whether the WHOIS details for a given domain suggest that the registrant is a business or an individual. Continue reading

DDoS-for-Hire Service Webstresser Dismantled

April 25, 2018

Authorities in the U.S., U.K. and the Netherlands on Tuesday took down popular online attack-for-hire service WebStresser.org and arrested its alleged administrators. Investigators say that prior to the takedown, the service had more than 136,000 registered users and was responsible for launching somewhere between four and six million attacks over the past three years.

The action, dubbed “Operation Power Off,” targeted WebStresser.org (previously Webstresser.co), one of the most active services for launching point-and-click distributed denial-of-service (DDoS) attacks. WebStresser was one of many so-called “booter” or “stresser” services — virtual hired muscle that anyone can rent to knock nearly any website or Internet user offline.

Webstresser.org (formerly Webstresser.co), as it appeared in 2017.

“The damage of these attacks is substantial,” reads a statement from the Dutch National Police in a Reddit thread about the takedown. “Victims are out of business for a period of time, and spend money on mitigation and on (other) security measures.”

In a separate statement released this morning, Europol — the law enforcement agency of the European Union — said “further measures were taken against the top users of this marketplace in the Netherlands, Italy, Spain, Croatia, the United Kingdom, Australia, Canada and Hong Kong.” The servers powering WebStresser were located in Germany, the Netherlands and the United States, according to Europol.

The U.K.’s National Crime Agency said WebStresser could be rented for as little as $14.99, and that the service allowed people with little or no technical knowledge to launch crippling DDoS attacks around the world.

Neither the Dutch nor U.K. authorities would say who was arrested in connection with this takedown. But according to information obtained by KrebsOnSecurity, the administrator of WebStresser allegedly was a 19-year-old from Prokuplje, Serbia named Jovan Mirkovic.

Mirkovic, who went by the hacker nickname “m1rk,” also used the alias “Mirkovik Babs” on Facebook where for years he openly discussed his role in programming and ultimately running WebStresser. The last post on Mirkovic’s Facebook page, dated April 3 (the day before the takedown), shows the young hacker sipping what appears to be liquor while bathing. Below that image are dozens of comments left in the past few hours, most of them simply, “RIP.”

Continue reading

Transcription Service Leaked Medical Records

April 23, 2018

MEDantex, a Kansas-based company that provides medical transcription services for hospitals, clinics and private physicians, took down its customer Web portal last week after being notified by KrebsOnSecurity that it was leaking sensitive patient medical records — apparently for thousands of physicians.

On Friday, KrebsOnSecurity learned that the portion of MEDantex’s site which was supposed to be a password-protected portal physicians could use to upload audio-recorded notes about their patients was instead completely open to the Internet.

What’s more, numerous online tools intended for use by MEDantex employees were exposed to anyone with a Web browser, including pages that allowed visitors to add or delete users, and to search for patient records by physician or patient name. No authentication was required to access any of these pages.

This exposed administrative page from MEDantex’s site granted anyone complete access to physician files, as well as the ability to add and delete authorized users.

Several MEDantex portal pages left exposed to the Web suggest that the company recently was the victim of WhiteRose, a strain of ransomware that encrypts a victim’s files unless and until a ransom demand is paid — usually in the form of some virtual currency such as bitcoin.

Contacted by KrebsOnSecurity, MEDantex founder and chief executive Sreeram Pydah confirmed that the Wichita, Kansas based transcription firm recently rebuilt its online servers after suffering a ransomware infestation. Pydah said the MEDantex portal was taken down for nearly two weeks, and that it appears the glitch exposing patient records to the Web was somehow incorporated into that rebuild.

“There was some ransomware injection [into the site], and we rebuilt it,” Pydah said, just minutes before disabling the portal (which remains down as of this publication). “I don’t know how they left the documents in the open like that. We’re going to take the site down and try to figure out how this happened.”

It’s unclear exactly how many patient records were left exposed on MEDantex’s site. But one of the main exposed directories was named “/documents/userdoc,” and it included more than 2,300 physicians listed alphabetically by first initial and last name. Drilling down into each of these directories revealed a varying number of patient records — displayed and downloadable as Microsoft Word documents and/or raw audio files.

Although many of the exposed documents appear to be quite recent, some of the records dated as far back as 2007. It’s also unclear how long the data was accessible, but this Google cache of the MEDantex physician portal seems to indicate it was wide open on April 10, 2018.

Among the clients listed on MEDantex’s site include New York University Medical Center; San Francisco Multi-Specialty Medical Group; Jackson Hospital in Montgomery Ala.; Allen County Hospital in Iola, Kan; Green Clinic Surgical Hospital in Ruston, La.; Trillium Specialty Hospital in Mesa and Sun City, Ariz.; Cooper University Hospital in Camden, N.J.; Sunrise Medical Group in Miami; the Wichita Clinic in Wichita, Kan.; the Kansas Spine Center; the Kansas Orthopedic Center; and Foundation Surgical Hospitals nationwide. MEDantex’s site states these are just some of the healthcare organizations partnering with the company for transcription services. Continue reading

Is Facebook’s Anti-Abuse System Broken?

April 20, 2018

Facebook has built some of the most advanced algorithms for tracking users, but when it comes to acting on user abuse reports about Facebook groups and content that clearly violate the company’s “community standards,” the social media giant’s technology appears to be woefully inadequate.

Last week, Facebook deleted almost 120 groups totaling more than 300,000 members. The groups were mostly closed — requiring approval from group administrators before outsiders could view the day-to-day postings of group members.

However, the titles, images and postings available on each group’s front page left little doubt about their true purpose: Selling everything from stolen credit cards, identities and hacked accounts to services that help automate things like spamming, phishing and denial-of-service attacks for hire.

To its credit, Facebook deleted the groups within just a few hours of KrebsOnSecurity sharing via email a spreadsheet detailing each group, which concluded that the average length of time the groups had been active on Facebook was two years. But I suspect that the company took this extraordinary step mainly because I informed them that I intended to write about the proliferation of cybercrime-based groups on Facebook.

That story, Deleted Facebook Cybercrime Groups had 300,000 Members, ended with a statement from Facebook promising to crack down on such activity and instructing users on how to report groups that violate it its community standards.

In short order, some of the groups I reported that were removed re-established themselves within hours of Facebook’s action. I decided instead of contacting Facebook’s public relations arm directly that I would report those resurrected groups and others using Facebook’s stated process. Roughly two days later I received a series replies saying that Facebook had reviewed my reports but that none of the groups were found to have violated its standards. Here’s a snippet from those replies:

Perhaps I should give Facebook the benefit of the doubt: Maybe my multiple reports one after the other triggered some kind of anti-abuse feature that is designed to throttle those who would seek to abuse it to get otherwise legitimate groups taken offline — much in the way that pools of automated bot accounts have been known to abuse Twitter’s reporting system to successfully sideline accounts of specific targets.

Or it could be that I simply didn’t click the proper sequence of buttons when reporting these groups. The closest match I could find in Facebook’s abuse reporting system were, “Doesn’t belong on Facebook,” and “Purchase or sale of drugs, guns or regulated products.” There was/is no option for “selling hacked accounts, credit cards and identities,” or anything of that sort.

In any case, one thing seems clear: Naming and shaming these shady Facebook groups via Twitter seems to work better right now for getting them removed from Facebook than using Facebook’s own formal abuse reporting process. So that’s what I did on Thursday. Here’s an example: Continue reading