As a researcher in Usable Privacy and Human-Computer Interaction Maximiliane Windl explores how privacy and data protection mechanisms can be designed to keep pace with technological advances while remaining convenient and easy to use. Her work combines technical solutions with an understanding of how people actually perceive and use systems. She sees artificial intelligence not only as a challenge to privacy, but also as a potential part of the solution. Her goal is to develop intelligent systems that help people understand, manage, and control their data. Among other things, she is exploring household robots, augmented reality, and emerging sensor technologies.
Windl also investigates how abstract privacy decisions can be made tangible and directly understandable. “With physical solutions, I don’t have to trust that some algorithm in the background is doing the right thing. If I slide a cover over a camera, for example, I can see that it is blocked and can’t record me. That’s it.” Windle applied this principle to the smart home in one research project, developing a physical dashboard with controls for connected devices and comparing its use with an app. “The participants actually found the dashboard faster and easier to use. They could simply turn a dial instead of first taking out their smartphone and opening the app.”
Privacy Shouldn’t Become a Constant Chore
“The basic problem with privacy is that most people don’t feel like dealing with it all the time. Who actually reads a privacy policy or a cookie banner? We really just want these things to go away because they get in the way of whatever we’re trying to do.” A privacy assistant could know a person’s privacy preferences and make everyday decisions on their behalf, rather than repeatedly confronting them with consent dialogs and settings.
“The idea of a personalized privacy assistant isn’t new. People have been thinking about it for around ten years. But I believe it is more realistic today than ever before to actually build such a system.” According to Windl, one of the central challenges with earlier approaches was that static, rule-based systems could not keep up with the complexity and scale of today’s technologies. Recent advances in artificial intelligence, particularly large language models and adaptive learning, could open up new possibilities for addressing these limitations. “If an AI assistant makes privacy decisions for me, we obviously need to figure out how to establish trust in that system. People still need to remain in control: They need to be able to change decisions, understand what the system has done, and intervene at any time.”
Understanding Technology to Better Assess Its Risks
To develop effective privacy mechanisms, Windl considers technological capabilities and human perception together: What can devices and systems actually do, and how do people perceive their capabilities and risks? “We need to think about privacy from the very beginning of the development process, rather than only afterward,” the researcher says. That is why she deliberately looks beyond her own field and attends technical conferences to observe developments that could raise new questions about privacy and data protection.
At the same time, Windl investigates people’s privacy needs and how those needs can be reliably identified. Going forward, she wants to explore new methods that go beyond self-reporting: “When we ask people about their behavior in interviews or questionnaires, they often respond in ways that reflect how they would like to behave themselves. But when we actually observe them using a system, we sometimes find that they behave very differently.”
People also do not always intuitively assess the capabilities of technological systems correctly. According to the researcher, humanoid robots provide a particularly striking example: “They have eyes, so people intuitively assume that they can see through those eyes as well. But in reality, they may have completely different sensors built into them. They look like humans, but their perception may work nothing like a human’s.” These sensors can include radio-frequency sensors that detect information about movement, in some cases even through walls. According to Windl, people often underestimate what kind of sensitive information can be inferred from such sensor data.
Advancing Privacy Alongside Technology
Windl, who joined CISPA after moving from Ludwig Maximilian University of Munich, is convinced that research can drive change: “Especially as more and more intelligent devices enter our homes and become capable of collecting increasingly more information about us, we shouldn’t simply stand by and watch.” At CISPA, she therefore plans to work on advancing privacy mechanisms alongside new technologies, bringing together technical capabilities and human needs.
