Entries Tagged "anonymity"

Page 3 of 8

Apple's Differential Privacy

At the Apple Worldwide Developers Conference earlier this week, Apple talked about something called “differential privacy.” We know very little about the details, but it seems to be an anonymization technique designed to collect user data without revealing personal information.

What we know about anonymization is that it’s much harder than people think, and it’s likely that this technique will be full of privacy vulnerabilities. (See, for example, the excellent work of Latanya Sweeney.) As expected, security experts are skeptical. Here’s Matt Green trying to figure it out.

So while I applaud Apple for trying to improve privacy within its business models, I would like some more transparency and some more public scrutiny.

EDITED TO ADD (6/17): Adam Shostack comments. And more commentary from Tom’s Guide.

EDITED TO ADD (6/17): Here’s a slide deck on privacy from the WWDC.

Posted on June 16, 2016 at 9:30 PM • View Comments

Identifying People from Their Metadata

Jonathan Mayer, Patrick Mutchler, and John C. Mitchell, “Evaluating the privacy properties of telephone metadata“:

Abstract: Since 2013, a stream of disclosures has prompted reconsideration of surveillance law and policy. One of the most controversial principles, both in the United States and abroad, is that communications metadata receives substantially less protection than communications content. Several nations currently collect telephone metadata in bulk, including on their own citizens. In this paper, we attempt to shed light on the privacy properties of telephone metadata. Using a crowdsourcing methodology, we demonstrate that telephone metadata is densely interconnected, can trivially be reidentified, and can be used to draw sensitive inferences.

New research, but not a new result. There have been several similar studies over the years. This one uses only anonymized call and SMS metadata to identify people who volunteered for the study.

Four assorted news articles.

Posted on May 19, 2016 at 6:10 AM • View Comments

Tracking Anonymous Web Users

This research shows how to track e-commerce users better across multiple sessions, even when they do not provide unique identifiers such as user IDs or cookies.

Abstract: Targeting individual consumers has become a hallmark of direct and digital marketing, particularly as it has become easier to identify customers as they interact repeatedly with a company. However, across a wide variety of contexts and tracking technologies, companies find that customers can not be consistently identified which leads to a substantial fraction of anonymous visits in any CRM database. We develop a Bayesian imputation approach that allows us to probabilistically assign anonymous sessions to users, while ac- counting for a customer’s demographic information, frequency of interaction with the firm, and activities the customer engages in. Our approach simultaneously estimates a hierarchical model of customer behavior while probabilistically imputing which customers made the anonymous visits. We present both synthetic and real data studies that demonstrate our approach makes more accurate inference about individual customers’ preferences and responsiveness to marketing, relative to common approaches to anonymous visits: nearest- neighbor matching or ignoring the anonymous visits. We show how companies who use the proposed method will be better able to target individual customers, as well as infer how many of the anonymous visits are made by new customers.

Posted on February 5, 2016 at 6:56 AM • View Comments

De-Anonymizing Users from their Coding Styles

Interesting blog post:

We are able to de-anonymize executable binaries of 20 programmers with 96% correct classification accuracy. In the de-anonymization process, the machine learning classifier trains on 8 executable binaries for each programmer to generate numeric representations of their coding styles. Such a high accuracy with this small amount of training data has not been reached in previous attempts. After scaling up the approach by increasing the dataset size, we de-anonymize 600 programmers with 52% accuracy. There has been no previous attempt to de-anonymize such a large binary dataset. The abovementioned executable binaries are compiled without any compiler optimizations, which are options to make binaries smaller and faster while transforming the source code more than plain compilation. As a result, compiler optimizations further normalize authorial style. For the first time in programmer de-anonymization, we show that we can still identify programmers of optimized executable binaries. While we can de-anonymize 100 programmers from unoptimized executable binaries with 78% accuracy, we can de-anonymize them from optimized executable binaries with 64% accuracy. We also show that stripping and removing symbol information from the executable binaries reduces the accuracy to 66%, which is a surprisingly small drop. This suggests that coding style survives complicated transformations.

Here’s the paper.

And here’s their previous paper, de-anonymizing programmers from their source code.

Posted on January 4, 2016 at 7:41 AM • View Comments

Did Carnegie Mellon Attack Tor for the FBI?

There’s pretty strong evidence that the team of researchers from Carnegie Mellon University who cancelled their scheduled 2015 Black Hat talk deanonymized Tor users for the FBI.

Details are in this Vice story and this Wired story (and these two follow-on Vice stories). And here’s the reaction from the Tor Project.

Nicholas Weaver guessed this back in January.

The behavior of the researchers is reprehensible, but the real issue is that CERT Coordination Center (CERT/CC) has lost its credibility as an honest broker. The researchers discovered this vulnerability and submitted it to CERT. Neither the researchers nor CERT disclosed this vulnerability to the Tor Project. Instead, the researchers apparently used this vulnerability to deanonymize a large number of hidden service visitors and provide the information to the FBI.

Does anyone still trust CERT to behave in the Internet’s best interests?

EDITED TO ADD (12/14): I was wrong. CERT did disclose to Tor.

Posted on November 16, 2015 at 6:19 AM • View Comments

Anonymous Browsing at the Library

A rural New Hampshire library decided to install Tor on their computers and allow anonymous Internet browsing. The Department of Homeland pressured them to stop:

A special agent in a Boston DHS office forwarded the article to the New Hampshire police, who forwarded it to a sergeant at the Lebanon Police Department.

DHS spokesman Shawn Neudauer said the agent was simply providing “visibility/situational awareness,” and did not have any direct contact with the Lebanon police or library. “The use of a Tor browser is not, in [or] of itself, illegal and there are legitimate purposes for its use,” Neudauer said, “However, the protections that Tor offers can be attractive to criminal enterprises or actors and HSI [Homeland Security Investigations] will continue to pursue those individuals who seek to use the anonymizing technology to further their illicit activity.”

When the DHS inquiry was brought to his attention, Lt. Matthew Isham of the Lebanon Police Department was concerned. “For all the good that a Tor may allow as far as speech, there is also the criminal side that would take advantage of that as well,” Isham said. “We felt we needed to make the city aware of it.”

The good news is that the library is resisting the pressure and keeping Tor running.

This is an important issue for reasons that go beyond the New Hampshire library. The goal of the Library Freedom Project is to set up Tor exit nodes at libraries. Exit nodes help every Tor user in the world; the more of them there are, the harder it is to subvert the system. The Kilton Public Library isn’t just allowing its patrons to browse the Internet anonymously; it is helping dissidents around the world stay alive.

Librarians have been protecting our privacy for decades, and I’m happy to see that tradition continue.

EDITED TO ADD (10/13): As a result of the story, more libraries are planning to run Tor nodes.

Posted on September 16, 2015 at 1:40 PM • View Comments

Organizational Doxing of Ashley Madison

The—depending on who is doing the reporting—cheating, affair, adultery, or infidelity site Ashley Madison has been hacked. The hackers are threatening to expose all of the company’s documents, including internal e-mails and details of its 37 million customers. Brian Krebs writes about the hackers’ demands.

According to the hackers, although the “full delete” feature that Ashley Madison advertises promises “removal of site usage history and personally identifiable information from the site,” users’ purchase details—including real name and address—aren’t actually scrubbed.

“Full Delete netted ALM $1.7mm in revenue in 2014. It’s also a complete lie,” the hacking group wrote. “Users almost always pay with credit card; their purchase details are not removed as promised, and include real name and address, which is of course the most important information the users want removed.”

Their demands continue:

“Avid Life Media has been instructed to take Ashley Madison and Established Men offline permanently in all forms, or we will release all customer records, including profiles with all the customers’ secret sexual fantasies and matching credit card transactions, real names and addresses, and employee documents and emails. The other websites may stay online.”

Established Men is another of the company’s sites; this one is designed to link wealthy men with young and pretty women.

This is yet another instance of organizational doxing:

Dumping an organization’s secret information is going to become increasingly common as individuals realize its effectiveness for whistleblowing and revenge. While some hackers will use journalists to separate the news stories from mere personal information, not all will.

EDITED TO ADD (7/22): I don’t believe they have 37 million users. This type of service will only appeal to a certain socio-economic demographic, and it’s not equivalent to 10% of the US population.

This page claims that 20% of the population of Ottawa is registered. Given that 25% of the population are children, that means it’s 30% of the adult population: 189,000 people. I just don’t believe it.

Posted on July 20, 2015 at 3:15 PM • View Comments

Why We Encrypt

Encryption protects our data. It protects our data when it’s sitting on our computers and in data centers, and it protects it when it’s being transmitted around the Internet. It protects our conversations, whether video, voice, or text. It protects our privacy. It protects our anonymity. And sometimes, it protects our lives.

This protection is important for everyone. It’s easy to see how encryption protects journalists, human rights defenders, and political activists in authoritarian countries. But encryption protects the rest of us as well. It protects our data from criminals. It protects it from competitors, neighbors, and family members. It protects it from malicious attackers, and it protects it from accidents.

Encryption works best if it’s ubiquitous and automatic. The two forms of encryption you use most often—https URLs on your browser, and the handset-to-tower link for your cell phone calls—work so well because you don’t even know they’re there.

Encryption should be enabled for everything by default, not a feature you turn on only if you’re doing something you consider worth protecting.

This is important. If we only use encryption when we’re working with important data, then encryption signals that data’s importance. If only dissidents use encryption in a country, that country’s authorities have an easy way of identifying them. But if everyone uses it all of the time, encryption ceases to be a signal. No one can distinguish simple chatting from deeply private conversation. The government can’t tell the dissidents from the rest of the population. Every time you use encryption, you’re protecting someone who needs to use it to stay alive.

It’s important to remember that encryption doesn’t magically convey security. There are many ways to get encryption wrong, and we regularly see them in the headlines. Encryption doesn’t protect your computer or phone from being hacked, and it can’t protect metadata, such as e-mail addresses that need to be unencrypted so your mail can be delivered.

But encryption is the most important privacy-preserving technology we have, and one that is uniquely suited to protect against bulk surveillance—the kind done by governments looking to control their populations and criminals looking for vulnerable victims. By forcing both to target their attacks against individuals, we protect society.

Today, we are seeing government pushback against encryption. Many countries, from States like China and Russia to more democratic governments like the United States and the United Kingdom, are either talking about or implementing policies that limit strong encryption. This is dangerous, because it’s technically impossible, and the attempt will cause incredible damage to the security of the Internet.

There are two morals to all of this. One, we should push companies to offer encryption to everyone, by default. And two, we should resist demands from governments to weaken encryption. Any weakening, even in the name of legitimate law enforcement, puts us all at risk. Even though criminals benefit from strong encryption, we’re all much more secure when we all have strong encryption.

This originally appeared in Securing Safe Spaces Online.

EDITED TO ADD: Last month, I blogged about a UN report on the value of encryption technologies to human freedom worldwide. This essay is the foreword to a companion document:

To support the findings contained in the Special Rapporteur’s report, Privacy International, the Harvard Law School’s International Human Rights Law Clinic and ARTICLE 19 have published an accompanying booklet, Securing Safe Spaces Online: Encryption, online anonymity and human rights which explores the impact of measures to restrict online encryption and anonymity in four particular countries ­—the United Kingdom, Morocco, Pakistan and South Korea.

EDITED TO ADD (7/8): this essay has been translated into Russian.

EDITED TO ADD (4/29/2019): this essay has been translated into Spanish.

EDITED TO ADD (11/13/2019): this essay has been translated into Bosnian.

EDITED TO ADD (10/28/2020): this essay has been translated into French and Hungarian.

EDITED TO ADD: this essay has been translated into German.

Posted on June 23, 2015 at 6:02 AM • View Comments

Eighth Movie-Plot Threat Contest Winner

On April 1, I announced the Eighth Movie-Plot Threat Contest:

I want a movie-plot threat that shows the evils of encryption. (For those who don’t know, a movie-plot threat is a scary-threat story that would make a great movie, but is much too specific to build security policies around. Contest history here.) We’ve long heard about the evils of the Four Horsemen of the Internet Apocalypse—terrorists, drug dealers, kidnappers, and child pornographers. (Or maybe they’re terrorists, pedophiles, drug dealers, and money launderers; I can never remember.) Try to be more original than that. And nothing too science fictional; today’s technology or presumed technology only.

On May 14, I announced the five semifinalists. The votes are in, and the winner is TonyK:

November 6 2020, the morning of the presidential election. This will be the first election where votes can be cast from smart phones and laptops. A record turnout is expected.

There is much excitement as live results are being displayed all over the place. Twitter, television, apps and websites are all displaying the vote counts. It is a close race between the leading candidates until about 9 am when a third candidate starts to rapidly close the gap. He was an unknown independent that had suspected ties to multiple terrorist organizations. There was outrage when he got on to the ballot, but it had quickly died down when he put forth no campaign effort.

By 11 am the independent was predicted to win, and the software called it for him at 3:22 pm.

At 4 the CEO of the software maker was being interviewed on CNN. There were accusations of everything from bribery to bugs to hackers being responsible for the results. Demands were made for audits and recounts. Some were even asking for the data to be made publicly available. The CEO calmly explained that there could be no audit or recount. The system was encrypted end to end and all the votes were cryptographically anonymized.

The interviewer was stunned and sat there in silence. When he eventually spoke, he said “We just elected a terrorist as the President of the United States.”

For the record, Nick P was a close runner-up.

Congratulations, TonyK. Contact me by e-mail, and I’ll send you your fabulous prizes.

Previous contests.

EDITED TO ADD (6/14): Slashdot thread.

Posted on June 13, 2015 at 12:11 PM • View Comments

1 2 3 4 5 … 8

Sidebar photo of Bruce Schneier by Joe MacInnis.