Apu Kapadia, an associate professor of computer science at the School of Informatics, Computing, and Engineering, is part of a collaboration that has earned a grant from the National Science Foundation to study user-centric designs for sensor-based technology in the home that can assure privacy.
The increased use of digital assistants and high-fidelity, high-definition cameras, microphones, and sensors in the home has increased home security, improved the quality of life for the aging and disabled, and made the use of cloud services for shopping, music, and more easier than ever. However, such convenience comes with a privacy tradeoff. Voice-activated devices record casual conversations or encounters, and that information may be captured, disseminated, and archived without a person’s knowledge. Often people cannot tell whether a camera or microphone is actually ‘on’ or ‘off’. Kapadia, through his collaboration with researchers from the University of Pittsburgh, hopes to explore novel physical designs of sensors that convey a “tangible” sense of privacy.
“The simplest example of ‘tangible privacy’ that we are all familiar with is a physical lens cap that obscures a camera lens,” Kapadia said. “Absent such mechanisms, people find workarounds such as placing stickers on their laptop cameras, for example. But it is hard to use similar workarounds for cameras that don't belong to you or that are out of reach.”
Few current devices feature a clear indicator of when an audio or video sensor is engaged, and some sensors have been hacked and used to record people without their consent or knowledge. Digital assistant devices, such as Amazon’s Echo, are constantly listening to sense a specific trigger word that makes a device “wake up” to serve the user, and that listening function has proven to be problematic on occasion. Hackers have shown an ability to take control of such devices, allowing outsiders to intrude on people’s privacy.
Kapadia and his group will study how people interpret current designs as well as the shortcomings of current IoT devices in communicating to people nearby what is being recorded. They also will explore sensor designs that are more intuitive and trustworthy.
“We hope that this work will greatly improve people's privacy—and sense of privacy—in this increasingly networked age where we're surrounded by IoT devices,” Kapadia said. “With so many sensors around us, it is important to know when one is being recorded and, for example, when one's conversations are being sent off to the cloud.”
The NSF grant is worth $450,000 in total, with IU receiving $160,000.
“The British parliament found it funny when the iPhone belonging to Britain’s defense secretary interrupted his speech when it mistook ‘Syria’ for ‘Siri,’” said Amr Sabry, the chair of computer science at SICE. “Perhaps this is superficially amusing, but it is ominous that practically every word anybody says is processed, recorded, and almost certainly stored. This happens during the most private, intimate, and sensitive conversations. Apu’s project addresses this widespread problem and promises to investigate tangible solutions that allow users some control over their privacy.”
For more information on security at SICE, visit our website.