Privacy-focused News Recommendation

Welcome to the usage study

Existing news recommendation systems collect all user data in order to provide personalized recommendations. In order to address the privacy concerns, we are researching ways to personalize news completely locally on the user's laptop. We call our local news recommendation system PrIA . This study is approved by the IRB at Stony Brook.

During the user study, we will present you with news recommendations. Some of the news recommendations are from our local PrIA system and some of the recommendations are those provided by Google's news recommendation service. The recommendations will be randomized. We will ask you to tell us which recommendations are relevant.

How can I participate?

To participate, you should be at least 18 years of age, have an Android mobile phone(with chrome installed), have either of Windows/Mac machine(with chrome installed).

If you agree to take part in this study, you need to register with a user name. You will then need to,

1. Download and install our tool on your mobile phone and your personal computer. Go to downloads.

2. Read and sign the consent form.

3. You will be asked to provide a username. You can provide any username. You will also be asked to provide a password (that is not transferred to us). This password will be used to encrypt any data that is transferred between your mobile and your laptop.

4. Recommendations pop up on your phone. Swipe right if you find the recommendation relevant, left if you dont.

What data on my phone does the study access?

The software works on the following pieces of data on your mobile phone:
    a. Browser history.
    b. Facebook and Twitter feeds.

None of this data will be stored externally, and will only be stored at the user's mobile and the user's laptop. We will move the encrypted data from your mobile phone to your laptop through our server. The encrypted data will be deleted immediately from the server after the transfer or after a short time duration if the transfer fails.

The data is encrypted using the password provided by the user. The password is not known to the researchers.

Researchers will only have access to the user ratings of the recommendations. The actual recommendations (URL or the content) will not be visible. The researchers will only have information such as "Out of the 30 recommendations, user X finds recommendations 1, 4, and 5 to be relevant.". The content of the recommendation is not available. However, the researchers do know which of the recommendations were generated by PrIA versus Google's cloud-based news recommendation.

Our code is open source and open for inspection.

What are my risks?

We have made every effort to minimize your risks in participating in this study.

1. No personally identifiable information is logged. This includes your phone number, IMEI number, IP address etc

2. Researchers will not look at the user's data. Recommendations are done purely by our machine algorithm.

3. Researchers would only look at the feedback given by user. For example, the researchers will have the following information: "Out of the 30 recommended articles, the user liked articles 1, 5, and 9". However, the content of the articles will not be available to the researchers. The server also has information about which of the recommendations were PrIA recommended and which were generated by Google’s recommendation system

4. You can stop participating in the study at any time and request that all data collected from your device be removed

5. The study server is password protected and only the study researchers can access this information

How will this study benefit me?

We don't have any monetary benefits for you. However, feel free to bask in the glory of having helped some research work.

How can I withdraw from the study?

You can withdraw from the study at any time. Please go to the uninstall page for more details.