“Is Our Children’s Apps Learning?” Automatically Detecting COPPA Violations
نویسندگان
چکیده
In recent years, a market of games and learning apps for children has flourished in the mobile world. Many of these often “free” mobile apps have access to a variety of sensitive personal information about the user, which app developers can monetize via advertising or other means. In the United States, the Children’s Online Privacy Protection Act (COPPA) protects children’s privacy, requiring parental consent to the use of personal information and prohibiting behavioral advertising and online tracking. In this work, we present our ongoing effort to develop a method to automatically evaluate mobile apps’ COPPA compliance. Our method combines dynamic execution analysis (to track sensitive resource access at runtime) with traffic monitoring (to reveal private information leaving the device and recording with whom it gets shared, even if encrypted). We complement empirical technical observations with legal analysis of the apps’ corresponding privacy policies. As a proof of concept, we scraped the Google Play store for apps distributed in categories specifically targeting users under than 13 years of age, which subjects these products to COPPA’s regulations. We automated app execution on an instrumented version of the Android OS, recording the apps’ access to and transmission of sensitive information. To contextualize third parties (e.g., advertising networks) with whom the apps share information, we leveraged a crowdsourced dataset collected by the Lumen Privacy Tool (formerly Haystack) [27], an Androidbased device-local traffic inspection platform. Our effort seeks to illuminate apps’ compliance with COPPA and catalog the organizations that collect sensitive user information. In our preliminary results, we find several likely COPPA violations, including omission of prior consent and active sharing of persistent identifiers with third-party services for tracking and profiling of children. These results demonstrate our testbed’s capability to detect different types of possible violations in the market for children’s apps.
منابع مشابه
Brahmastra: Driving Apps to Test the Security of Third-Party Components
We present an app automation tool called Brahmastra for helping app stores and security researchers to test thirdparty components in mobile apps at runtime. The main challenge is that call sites that invoke third-party code may be deeply embedded in the app, beyond the reach of traditional GUI testing tools. Our approach uses static analysis to construct a page transition graph and discover exe...
متن کاملOn Providing Automatic Parental Consent over Information Collection from Children
Children’s privacy has become critical with the increasing use of the Internet for commercial purposes and corresponding increase in requests for information. 65% of children between the ages of 10 and 13 use the Internet for casual web surfing, chatting, games, schoolwork, e-mail, interactive learning, and other applications. Often, websites hosting these activities ask for personal informatio...
متن کاملAn examination of children's learning progression shifts while using touch screen virtual manipulative mathematics apps
The purpose of this study was to examine shifts in young children’s learning progression levels while they interacted with virtual manipulative mathematics apps on touch-screen devices. A total of 100 children participated in six mathematics learning sequences while using 18 virtual manipulative mathematics touch-screen apps during clinical interviews. Researchers developed a micro-scoring tool...
متن کاملAutomating Privacy Testing of Smartphone Applications
Smartphones have revolutionized mobile computing, but have created concerns that many third-party mobile applications do not properly handle users’ privacy-sensitive data. In this paper, we propose AppInspector, an automated privacy validation system that analyzes apps and generates reports of potential privacy risks. A key insight is that distinguishing acceptable disclosures from privacy viol...
متن کاملBoostClean: Automated Error Detection and Repair for Machine Learning
Predictive models based on machine learning can be highly sensitive to data error. Training data are often combined from a variety of different sources, each susceptible to different types of inconsistencies, and as new data stream in during prediction time, the model may encounter previously unseen inconsistencies. An important class of such inconsistencies are domain value violations that occ...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 2016