Artificial Intelligence, machine learning, and other technologies are changing the world in which we live and work in some subtle, and not-so-subtle, ways. And we’re diving into just a few of them in this podcast series produced in association with Darktrace.
One of the most visible places where analytics based on AI and machine learning are working their way into our popular awareness is in the realm of professional sports. From the virtual lines drawn on a football field to show the line of scrimmage and first-down markers, to Major League Baseball stat casts predicting the probability of successful base stealing, AI has become part of how we consume sports.
In this episode, Ars editors Sean Gallagher and Lee Hutchinson talked with Tim Wade, vice president at NTT’s Advanced Technology Group, about how NTT provides AI-based analytics for the Tour de France, the iconic 21-stage cycling competition. Wade’s team uses wireless communications, helicopters, and a data center in a truck to turn sensor data from each of the competitors’ bicycles into live statistics and analysis of the race—including an algorithm that predicts pile-ups in the peloton.
This special edition of the Ars Technicast podcast can be accessed in the following places:
iTunes: https://itunes.apple.com/us/podcast/the-ars-technicast/id522504024?mt=2 (Might take several hours after publication to appear.)
Yesterday, a security researcher who goes by illusionofchaos dropped public notice of three zero-day vulnerabilities in Apple’s iOS mobile operating system. The vulnerability disclosures are mixed in with the researcher’s frustration with Apple’s Security Bounty program, which illusionofchaos says chose to cover up an earlier-reported bug without giving them credit.
This researcher is by no means the first to publicly express their frustration with Apple over its security bounty program.
Nice bug—now shhh
illusionofchaos says that they’ve reported four iOS security vulnerabilities this year—the three zero-days they publicly disclosed yesterday plus an earlier bug that they say Apple fixed in iOS 14.7. It appears that their frustration largely comes from how Apple handled that first, now-fixed bug in analyticsd.
This now-fixed vulnerability allowed arbitrary user-installed apps to access iOS’s analytics data—the stuff that can be found in Settings --> Privacy --> Analytics & Improvements --> Analytics Data—without any permissions granted by the user. illusionofchaos found this particularly disturbing, because this data includes medical data harvested by Apple Watch, such as heart rate, irregular heart rhythm, atrial fibrillation detection, and so forth.
Analytics data was available to any application, even if the user disabled the iOS Share Analytics setting.
According to illusionofchaos, they sent Apple the first detailed report of this bug on April 29. Although Apple responded the next day, it did not respond to illusionofchaos again until June 3, when it said it planned to address the issue in iOS 14.7. On July 19, Apple did indeed fix the bug with iOS 14.7, but the security content list for iOS 14.7 acknowledged neither the researcher nor the vulnerability.
Apple told illusionofchaos that its failure to disclose the vulnerability and credit them was just a “processing issue” and that proper notice would be given in “an upcoming update.” The vulnerability and its resolution still were not acknowledged as of iOS 14.8 on September 13 or iOS 15.0 on September 20.
Frustration with this failure of Apple to live up to its own promises led illusionofchaos to first threaten, then publicly drop this week’s three zero-days. In illusionofchaos‘ own words: “Ten days ago I asked for an explanation and warned then that I would make my research public if I don’t receive an explanation. My request was ignored so I’m doing what I said I would.”
We do not have concrete timelines for illusionofchaos‘ disclosure of the three zero-days, or of Apple’s response to them—but illusionofchaos says the new disclosures still adhere to responsible guidelines: “Google Project Zero discloses vulnerabilities in 90 days after reporting them to vendor, ZDI – in 120. I have waited much longer, up to half a year in one case.”
New vulnerabilities: Gamed, nehelper enumerate, nehelper Wi-Fi
The zero-days illusionofchaos dropped yesterday can be used by user-installed apps to access data that those apps should not have or have not been granted access to. We’ve listed them below—along with links to illusionofchaos‘ Github repos with proof of concept code—in order of (our opinion of) their severity:
Gamed zero-day exposes Apple ID email and full name, exploitable Apple ID authentication tokens, and read access to Core Duet and Speed Dial databases
Nehelper Wi-Fi zero-day exposes Wi-Fi information to apps that have not been granted that access
Nehelper Enumerate zero-day exposes information about what apps are installed on the iOS device
The Gamed 0-day is obviously the most severe, since it both exposes Personal Identifiable Information (PII) and may be used in some cases to be able to perform actions at *.apple.com that would normally need to be either instigated by the iOS operating system itself, or by direct user interactions.
The Gamed zero-day’s read access to Core Duet and Speed Dial databases is also particularly troubling, since that access can be used to gain a pretty complete picture of the user’s entire set of interactions with others on the iOS device—who is in their contact list, who they’ve contacted (using both Apple and third-party applications) and when, and in some cases even file attachments to individual messages.
The Wi-Fi zero-day is next on the list, since unauthorized access to the iOS device’s Wi-Fi info might be used to track the user—or, possibly, learn the credentials necessary to access the user’s Wi-Fi network. The tracking is typically a more serious concern, since physical proximity is generally required to make Wi-Fi credentials themselves useful.
One interesting thing about the Wi-Fi zero-day is the simplicity of both the flaw and the method by which it can be exploited: “XPC endpoint com.apple.nehelper accepts user-supplied parameter sdk-version, and if its value is less than or equal to 524288, com.apple.developer.networking.wifi-info entitlement check is skipped.” In other words, all you need to do is claim to be using an older software development kit—and if so, your app gets to ignore the check that should disclose whether the user consented to access.
The Nehelper Enumerate zero-day appears to be the least damaging of the three. It simply allows an app to check whether another app is installed on the device by querying for the other app’s bundleID. We haven’t come up with a particularly scary use of this bug on its own, but a hypothetical malware app might leverage such a bug to determine whether a security or antivirus app is installed and then use that information to dynamically adapt its own behavior to better avoid detection.
Assuming illusionofchaos‘ description of their disclosure timeline is correct—that they’ve waited for longer than 30 days, and in one case 180 days, to publicly disclose these vulnerabilities—it’s hard to fault them for the drop. We do wish they had included full timelines for their interaction with Apple on all four vulnerabilities, rather than only the already-fixed one.
We can confirm that this frustration of researchers with Apple’s security bounty policies is by no means limited to this one pseudonymous researcher. Since Ars published a piece earlier this month about Apple’s slow and inconsistent response to security bounties, several researchers have contacted us privately to express their own frustration. In some cases, researchers included video clips demonstrating exploits of still-unfixed bugs.
We have reached out to Apple for comment, but we have yet to receive any response as of press time. We will update this story with any response from Apple as it arrives.
Security researcher Amit Serper of Guardicore discovered a severe flaw in Microsoft’s autodiscover—the protocol which allows automagical configuration of an email account with only the address and password required. The flaw allows attackers who purchase domains named “autodiscover”—for example autodiscover.com, or autodiscover.co.uk—to intercept the clear-text account credentials of users who are having network difficulty (or whose admins incorrectly configured DNS).
Guardicore purchased several such domains and operated them as proof-of-concept credential traps from April 16 to August 25 of this year:
A web server connected to these domains received hundreds of thousands of email credentials—many of which also double as Windows Active Directory domain credentials—in clear text. The credentials are sent from clients which request the URL /Autodiscover/autodiscover.xml, with an HTTP Basic authentication header which already includes the hapless user’s Base64-encoded credentials.
Three major flaws contribute to the overall vulnerability: the Autodiscover protocol’s “backoff and escalate” behavior when authentication fails, its failure to validate Autodiscover servers prior to giving up user credentials, and its willingness to use insecure mechanisms such as HTTP Basic in the first place.
Failing upward with autodiscover
The Autodiscover protocol’s real job is the simplification of account configuration—you can perhaps rely on a normal user to remember their email address and password, but decades of computing have taught us that asking them to remember and properly enter details like POP3 or IMAP4, TLS or SSL, TCP 465 or TCP 587, and the addresses of actual mail servers are several bridges too far.
The Autodiscover protocol allows normal users to configure their own email accounts without help, by storing all of the nonprivate portions of account configuration on publicly accessible servers. When you set up an Exchange account in Outlook, you feed it an email address and a password: for example, email@example.com with password Hunter2.
Armed with the user’s email address, Autodiscover sets about finding configuration information in a published XML document. It will try both HTTP and HTTPS connections, to the following URLs. (Note: contoso is a Microsoftism, representing an example domain name rather than any specific domain.)
So far, so good—we can reasonably assume that anyone allowed to place resources in either example.contoso.com or its Autodiscover subdomain has been granted explicit trust by the owner of example.contoso.com itself. Unfortunately, if these initial connection attempts fail, Autodiscover will back off and try to find resources at a higher-level domain.
In this case, Autodiscover’s next step would be to look for /Autodiscover/Autodiscover.xml on contoso.com itself, as well as Autodiscover.contoso.com. If this fails, Autodiscover fails upward yet again—this time sending email and password information to autodiscover.com itself.
This would be bad enough if Microsoft owned autodiscover.com—but the reality is considerably murkier. That domain was originally registered in 2002 and is currently owned by an unknown individual or organization using GoDaddy’s WHOIS privacy shield.
In the approximately four months Guardicore ran its test credential trap, it collected 96,671 unique sets of email username and passwords in clear text. These credentials came from a wide array of organizations—publicly traded companies, manufacturers, banks, power companies, and more.
Affected users don’t see HTTPS/TLS errors in Outlook—when the Autodiscover protocol fails up from Autodiscover.contoso.com.br to Autodiscover.com.br, the protection afforded by contoso‘s ownership of its own SSL cert vanishes. Whoever purchased Autodiscover.com.br—in this case, Guardicore—simply provides their own certificate, which satisfies TLS warnings despite not belonging to contoso at all.
In many cases, the Outlook or similar client will offer its user’s credentials initially in a more secure format, such as NTLM. Unfortunately, a simple HTTP 401 from the web server requesting HTTP Basic auth in its place is all that’s necessary—upon which the client using Autodiscover will comply (typically without error or warning to the user) and send the credentials in Base64 encoded plain text, completely readable by the web server answering the Autodiscover request.
The truly bad news here is that, from the general public’s perspective, there is no mitigation strategy for this Autodiscover bug. If your organization’s Autodiscover infrastructure is having a bad day, your client will “fail upward” as described, potentially exposing your credentials. This flaw has not yet been patched—according to Microsoft Senior Director Jeff Jones, Guardicore disclosed the flaw publicly prior to reporting it to Microsoft.
If you’re a network administrator, you can mitigate the issue by refusing DNS requests for Autodiscover domains—if every request to resolve a domain beginning in “Autodiscover” is blocked, the Autodiscover protocol won’t be able to leak credentials. Even then, you must be careful: you might be tempted to “block” such requests by returning 127.0.0.1, but this might allow a clever user to discover someone else’s email and/or Active Directory credentials, if they can trick the target into logging into the user’s PC.
If you’re an application developer, the fix is simpler: don’t implement the flawed part of the Autodiscover spec in the first place. If your application never attempts to authenticate against an “upstream” domain in the first place, it won’t leak your users’ credentials via Autodiscover.
For more technical detail, we highly recommend Guardicore’s own blog post as well as Microsoft’s own Autodiscover documentation.
The semiconductor chip shortage that has so vexed the auto industry looks set to continue for quite some time, according to a new industry survey. More than half of the companies that were surveyed by IPC said they expected the shortage to last until at least the second half of 2022. And right now, the chip shortage is being exacerbated by rising costs and a shortage of workers.
According to the survey, 80 percent of chip makers say that it’s become hard to find workers who have to be specially trained to handle the highly toxic compounds used in semiconductor manufacturing. The problem is worse in North America and in Asia, where more companies are reporting rising labor costs compared to those in Europe.
But only a third of Asian chip makers say they are finding it harder to find qualified workers, compared to 67 percent of North American companies and 63 percent of European companies. That may well explain why fewer Asian semiconductor companies (42 percent) are reporting increasing order backlogs, compared to 65 percent of North American and 60 percent of European companies.
Just under half (46 percent) said they were retraining their current workers to fill the gaps, and nearly as many (44 percent) said they were increasing wages to make the jobs more attractive. Other popular measures include more flexible hours and more training opportunities for workers.
Even more of the companies surveyed said that rising material costs were a problem, too—90 percent globally, with nearly as many suggesting that trend will continue for another six months at least. IPC says that chip makers’ profit margins are shrinking as a result.
That’s probably already being felt by some of their customers. According to a report by AlixPartners, the auto industry will lose out on $210 billion in revenue in 2021, forecasting a shortfall in production of 7.7 million vehicles worldwide. That’s got the US government’s attention, too. On Thursday, Commerce Secretary Gina Raimondo is meeting automakers and tech firms, as well as semiconductor companies, to see if the federal government can help.