Skip to main content
March 9, 2020

NSF CAREER award will help Kassem Fawaz improve data privacy

Written By: Jason Daley

Every day, almost everyone in the United States uses a service, visits a website or interacts with a device that collects confidential data about their location, interests or activities. But very few take the time to read those services’ overly complicated privacy statements or adjust settings to protect their data. That’s led to millions of people losing control over their online privacy.

Kassem Fawaz, an assistant professor of electrical and computer engineering at the University of Wisconsin-Madison, hopes that receiving a prestigious National Science Foundation CAREER award will help him to continue to create tools that allow technology consumers take their privacy back.

“Every time you ask someone if they care about their privacy, they say, ‘Yes.’ When you ask them what they have done to protect their privacy, the answer is, ‘Mostly nothing,’ he says. “That’s because they lack usable tools to do something, and if they have the tools, they don’t know how to use them.”

Over the next five years, Fawaz hopes to create a new set of intuitive, easily accessible tools that inform consumers and also let them do something about data collection and processing. “The main objective of this CAREER proposal is to bridge the gap between users and the privacy practices of the services, websites, companies, apps and devices with which they interact,” he says.

Fawaz has already made some headway in developing those tools. In 2018, he and collaborators from the University of Michigan and École Polytechnique Fédérale de Lausanne in Switzerland released an AI-powered service called Pribot. The natural language processing network analyzes the fine print and legalese in the often difficult-to-understand privacy statements of websites, translating them into more digestible language. That allows users to ask questions about the privacy policies—like, ‘Does this site collect my location data,’ or, ‘Does it sell my info to third parties’—which are answered by a friendly chatbot.

A companion service, called Polisis, uses graphics to display the type of information a website or service collects and how users can protect their data. So far, Pribot and Polisis have amassed a database of more than 30,000 privacy policies and continue to add more.

But Fawaz says those official privacy policies are only a small part of the entire privacy landscape. Some companies post privacy information in terms of service, license agreements, blog posts, FAQs and through other avenues. “This is where the actual research effort starts. Can we get all of that information, conceptualize it, understand that, and then present it in a more intuitive manner?” he says. “And can we go through the privacy settings of companies, analyze them, and also present them in a more usable manner?”

Eventually, he hopes his research will lead to privacy solutions that are part of people’s daily lives, like an interface that allows consumers to understand and adjust the privacy settings of sites they use without digging for them.

He’s also interested in creating a way to assist shoppers. “If you go to or visit Best Buy, your shopping decision is influenced by consumer ratings, costs, number of downloads, and other information,” he says. “What we’re thinking about is how we can include privacy as a dimension to help consumers make privacy-aware decisions when shopping.”

Fawaz says that public education is an important part of his CAREER proposal as well, especially since many people believe the data they share can’t be linked to them. “They think their data is not important,” he says.

But when he uses an education module to demonstrate how a simple trace using location data can reveal where they live, work and even grocery shop, it makes the issue more concrete. Cross referencing that with public records and other data can reveal even more personal information. “When I show them that, they start understanding and appreciating privacy problems better,” he says.