At an upcoming event, computer science professors at Louisiana State University (LSU) will present a new mechanism that uses artificial intelligence (AI) to help smartphones analyze how people grip them to determine if they are in the hands of their owners or not. “[We’ve developed] a media sound-based authentication method to protect smartphone notification privacy unobtrusively,” write the researchers in their paper. “[The mechanism] wisely hides or presents sensitive content by verifying who is holding the phone.”
Get a Grip
LSU’s Computer Science Assistant Professor Chen Wang, together with Ph.D. student Long Huang, have designed a novel authentication mechanism based on acoustic sensing. It uses sounds like notification tones to map and verify the user’s hand gripping the device. Explaining the mechanism in their paper, the researchers argue that because sounds are signals, they are absorbed, dampened, reflected, or refracted by the user’s hands. Their authentication mechanism captures the sounds and the vibrations using the smartphone’s microphone and accelerometer to generate spectrograms, which are then processed by an AI-based algorithm. This helps the system map how the individual’s contacting palm interferes with the signals, in essence creating a new kind of handgrip biometric. If there is a match, the verification is successful, and the system allows the notification previews to be displayed. In case it can’t find a match, the system only displays the total number of pending notifications, and not their actual content. “Moreover, because the smartphone sensors are all embedded on the same motherboard, we develop a cross-domain method to validate such hard-to-forge physical relationships among the mic, speaker and accelerometer,” note the researchers. This makes the system tamper-proof, further strengthening the security and privacy of the device.
Hold That Thought
However, while industry experts acknowledge the novelty of the mechanism, they aren’t impressed by its implementation and use case. “If the premise is that they do not want to interrupt user experience, then I believe this is doomed from the start,” Lecio de Paula Jr., VP of Data Protection at KnowBe4, told Lifewire over email. “The technology requires sound be on for it to function properly, but most people keep their phones on silent or vibrate.” Cyberspace attorney Sean Griffin is someone who has disabled audio notifications on their phone altogether. He’s also skeptical about the real-world use of the handgrip authentication mechanism. “I’m not sure I hold my phone in exactly the same way every single time I pick it up, so there is a potential for false negatives,” suggested Griffin. De Paula Jr. doesn’t think the technology seems practical in use, considering that many other variables are at play in the real world. One concern that strikes him is the room’s acoustics and the impact they’d have on the authentication performance. Bill Leddy, VP of Product at LoginID, thinks the blocking notifications use case, while interesting, is a bit too narrow to find any takers. “I doubt most people would download an app if it can be implemented at the app level, much less pay for such a feature. Adding it to the operating system seems like a stretch, but maybe [that’s a possibility],” Leddy told Lifewire via email. Given the concerns, de Paula Jr. thinks the novel authentication mechanism doesn’t seem like an upgrade over the current authentication techniques, especially since facial recognition data is typically stored locally on the device for authentication purposes, which reduces the privacy risk. Griffin agrees and is doubtful about the handgrip authentication mechanism making it to a smartphone in the near future. “Most [smartphone companies] have already determined the path they want to follow with authentication, and use of AI with facial recognition technology is leading the pack at the moment,” de Paula Jr. said.
