Privacy Violations Using Microtargeted Ads: A Case Study

By Aleksandra Korolova. 2010

196 citations recorded by the source corpus.

In this paper we propose a new class of attacks that exploit advertising systems offering micro targeting capabilities in order to breach user privacy. We study the advertising system offered by the world's largest online social network, Face book, and the risks that the design of the system poses to the privacy of its users. We propose, describe and provide experimental evidence of several novel approaches to exploiting the advertising system in order to obtain private user information. We communicated our findings to Face book on July 13, 2010, and received a prompt response. On July 20, 2010, Face book launched a change to their advertising system that made the kind of attacks we describe more difficult but not impossible to implement.

Subfields: Artificial Intelligence, Information Systems, Sociology and Political Science

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Information Systems

In Sociology and Political Science

Citations in the corpus

Differential Privacy for Growing Databases (2018, background): “Analysis of sensitive data without formal privacy guarantees has led to numerous privacy violations in practice, for tasks such as recommender systems [CKN + 11], targeted advertising [Kor10], data anonymization [NS08], and deep learning [HAPC17].”
BLENDER: Enabling Local Search with a Hybrid Differential Privacy Model (2017, background): “In the last decade, we have witnessed scores of adhoc approaches that have turned out to be inadequate for protecting privacy [33, 23].”
On formalizing fairness in prediction with machine learning (2017, background): “ the ubiquity of social media, there are many avenues for personal information including protected a−ributes to be leaked from the information readily available in the public domain. Indeed, Korolova [43] note that using data available on Facebook pro•les, one can infer user age, sexual orientation,”
Privacy via the Johnson-Lindenstrauss Transform (2012, background): “Even in a scenario where the data mining algorithm is run by the data collector itself (that is, parties A and B are the same), privacy breaches are possible if the result of data mining is to be published [19].”
Silent Listeners: The Evolution of Privacy and Disclosure on Facebook (2013, background): “Korolova found in 2010 that the microtargeting capabilities of Facebook advertising system could be exploited for privacy violations [35].”
Bias in algorithmic filtering and personalization (2013, background): “A Facebook advertiser can target users who live within 50 miles of San Francisco, are male, between 24 and 30 years old, single, interested in women, like skiing, have graduated from Harvard and work at Apple (Korolova 2010).”
Investigating sources of PII used in Facebook’s targeted advertising (2018, background): “Malicious and discriminatory advertising In 2011, Korolova [27] found that malicious advertisers could infer personally identifying information about users who click on an ad.”
Testing the Lipschitz Property over Product Distributions with Applications to Data Privacy (2013, background): “Since the data which gets shared is extremely rich in user information, it poses serious privacy concerns about the users’ information contained in the data [23,6].”
PriPeARL: A Framework for Privacy-Preserving Analytics and Reporting at LinkedIn (2018, methodology): “h communities (e.g., [2–4, 15, 21, 23, 27, 28, 31, 33, 34]), as well as on the limitations of simple anonymization techniques (e.g., [5, 30, 32]) and on privacy attacks and challenges in ad targeting [24]. Based on the lessons learned from the privacy literature, we decided to make use of the rigoro”
Protecting privacy when mining and sharing user data (2012, background): “But in the quest for improvement, the companies sometimes underestimate the e↵ects these data mining and sharing practices can have on violating the privacy of individuals [53, 109].”
Differentially Private Matrix Completion, Revisited (2017, background): “Naturally, personalization problems require collecting and analyzing sensitive customer data like their preferences for various items, which can lead to serious privacy breaches (Korolova, 2010; Narayanan & Shmatikov, 2010; Calandrino et al., 2011).”
A Practical Application of Differential Privacy to Personalized Online Advertising (2011, background): “Despite the restrictive definition, it has been demonstrated that differentially private analyses exist for a variety of tasks including the approximation of numerical functions (by adding carefully chosen random noise that conceals any single individual’s contribution) [7, 2, 19, 10].”