AI-powered radar can spy on phone calls from 10 feet, exposing new privacy risks
A team of computer science researchers at Penn State has developed a startling new way to eavesdrop on phone calls remotely by decoding subtle vibrations emitted by a cellphone's earpiece.
Using millimeter-wave radar combined with an AI speech recognition system, their setup can capture and transcribe conversations from up to 10 feet away with about 60% accuracy.
This breakthrough raises significant privacy concerns about the potential misuse of such emerging technologies.
The research builds on a 2022 project where the team achieved up to 83% accuracy in recognizing 10 predefined words using a similar approach.
The new work extends this capability to continuous speech transcription, though the accuracy is lower due to the complexity of decoding noisy radar data.
'When we talk on a cellphone, we tend to ignore the vibrations that come through the earpiece and cause the whole phone to vibrate,' said first author Suryoday Basak, a doctoral candidate in computer science.
'If we capture these same vibrations using remote radars and bring in machine learning to help us learn what is being said, using context clues, we can determine whole conversations. By understanding what is possible, we can help the public be aware of the potential risks.'
The team used a millimeter-wave radar sensor, the same technology employed in self-driving cars, motion detectors, and 5G wireless networks, to measure the tiny surface vibrations generated by speech played through a phone earpiece.
To interpret this noisy and low-quality data, they adapted Whisper, an open-source AI speech recognition model developed for clean audio, using a low-rank adaptation machine learning technique.
This method allowed them to retrain just 1 percent of Whisper's parameters specifically for radar data, improving transcription results without rebuilding the entire model from scratch.
Radar tech breakthrough
The experimental setup involved positioning the radar sensor about three meters (10 feet) away from the phone to capture the minute vibrations.
The data was then fed into the customized AI model, which produced transcriptions with around 60 percent accuracy over a vocabulary of up to 10,000 words.
While this is far from perfect, the researchers noted that even partial keyword matches could have serious security implications.
'The result was transcriptions of conversations, with an expectation of some errors, which was a marked improvement from our 2022 version, which outputs only a few words,' said co-author Mahanth Gowda, associate professor of computer science and engineering.
'But even picking up partial matches for speech, such as keywords, are useful in a security context.'
The team compared their approach to lip reading, which typically captures only 30% to 40% of spoken words but can still help people infer conversations when combined with context.
Similarly, the radar-AI system's output, though imperfect, can reveal sensitive information when supplemented with prior knowledge or manual correction.
Privacy risks amplified
Basak emphasized the potential privacy risks posed by this emerging technology.
'Similar to how lip readers can use limited information to interpret conversations, the output of our model combined with contextual information can allow us to infer parts of a phone conversation from a few meters away,' he said.
'The goal of our work was to explore whether these tools could potentially be used by bad actors to eavesdrop on phone conversations from a distance. Our findings suggest that this is technically feasible under certain conditions, and we hope this raises public awareness so people can be more mindful during sensitive calls.'
The U.S. National Science Foundation supported the research, and the team stressed that their experiments are intended to highlight possible vulnerabilities before malicious actors exploit them.
They envision future efforts to develop protective measures to secure personal conversations from this kind of remote surveillance.
As wireless technology and AI evolve rapidly, this study serves as a crucial warning: even the faintest vibrations from your everyday devices can potentially betray your most private words.
The study has been published in, published in the Proceedings of WiSec 2025: 18th ACM Conference on Security and Privacy in Wireless and Mobile Networks.
Solve the daily Crossword
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Yahoo
2 days ago
- Yahoo
Eating this surprising fruit might be the key to getting a better night's sleep
Eating one avocado a day might be the key to a better night's sleep. Adults who included the fruit -- and yes, it's a fruit -- in their daily diet over the course of just six months reported better quality sleep compared to people who consumed fewer than two avocados each month, researchers said. 'Sleep is emerging as a key lifestyle factor in heart health, and this study invites us to consider how nutrition — and foods like avocado — can play a role in improving it,' Dr. Kristina Petersen, an associate professor of nutritional sciences at Penn State University, explained. The recent findings on sleep were tied to avocado's beneficial nutrients including tryptophan, folate, and magnesium. An essential mineral, magnesium is involved in muscle contraction and relaxation. Tryptophan and folate are involved in producing melatonin, the hormone that regulates our sleep cycles. These conclusions were drawn following an assessment of 969 U.S. adults who had what is considered a larger waistline -- 35 inches or more in women and 40 inches or more in men. The study was supported by the Avocado Nutrition Center, although the agency had no role in data collection, analysis, or interpretation. The researchers had initially intended to focus on heart health alone, before participants reported sleep-related benefits. "This was a cardiovascular health trial, making the sleep benefits more credible since they emerged as unexpected secondary findings in a well-designed randomized controlled trial," Dr. John Saito, a spokesperson for the American Academy of Sleep Medicine and pulmonologist at Children's Hospital of Orange County, told Verywell this week. Eating avocados daily was also associated with a better diet and lower cholesterol levels, the researchers reported, building on similar research from 2022. Studies have previously found that disrupted sleep is linked to high levels of cholesterol: the waxy, fat-like substance that can lead to heart attack and stroke. Improved heart health can help people to get better sleep, and getting enough sleep can also improve heart health and slash the risk of disease, according to the Centers for Disease Control and Prevention. Avocado's healthy fat and fiber contributed to positive cardiovascular impacts. The healthy fat can also help to reduce cholesterol levels, and lower the risk for potentially life-threatening cardiac events. The fiber is tied to a healthy gut and reduced risk of death from any cause. Adults should consume at least 25 grams of fiber each day, according to Harvard Medical School. Of course, healthy fat is still fat, and avocados are packed with both nutrients and calories. An entire large avocado can add upward of 400 calories to your daily diet, the Cleveland Clinic says. But, the fruits are still considered to be healthy when consumed in moderation like being added to a morning smoothie or a salad at lunch. Half an avocado has more potassium than a banana, the Harvard T.H. Chan School of Public Health said. Humans need potassium to help fend off high blood pressure, which can result in kidney disease, eye damage, coronary artery disease, and other complications, if it's left untreated. While the findings of the new study cannot be generalized to all populations, Petersen noted that heart health is influenced by many factors, such as fitness and genetics. 'This is an encouraging step in expanding the science around avocados and the potential benefits of consumption,' she said.
Yahoo
2 days ago
- Yahoo
AI-powered radar can spy on phone calls from 10 feet, exposing new privacy risks
Believe it or not, your phone's tiniest vibrations can reveal your conversations — thanks to AI. A team of computer science researchers at Penn State has developed a startling new way to eavesdrop on phone calls remotely by decoding subtle vibrations emitted by a cellphone's earpiece. Using millimeter-wave radar combined with an AI speech recognition system, their setup can capture and transcribe conversations from up to 10 feet away with about 60% accuracy. This breakthrough raises significant privacy concerns about the potential misuse of such emerging technologies. The research builds on a 2022 project where the team achieved up to 83% accuracy in recognizing 10 predefined words using a similar approach. The new work extends this capability to continuous speech transcription, though the accuracy is lower due to the complexity of decoding noisy radar data. 'When we talk on a cellphone, we tend to ignore the vibrations that come through the earpiece and cause the whole phone to vibrate,' said first author Suryoday Basak, a doctoral candidate in computer science. 'If we capture these same vibrations using remote radars and bring in machine learning to help us learn what is being said, using context clues, we can determine whole conversations. By understanding what is possible, we can help the public be aware of the potential risks.' The team used a millimeter-wave radar sensor, the same technology employed in self-driving cars, motion detectors, and 5G wireless networks, to measure the tiny surface vibrations generated by speech played through a phone earpiece. To interpret this noisy and low-quality data, they adapted Whisper, an open-source AI speech recognition model developed for clean audio, using a low-rank adaptation machine learning technique. This method allowed them to retrain just 1 percent of Whisper's parameters specifically for radar data, improving transcription results without rebuilding the entire model from scratch. Radar tech breakthrough The experimental setup involved positioning the radar sensor about three meters (10 feet) away from the phone to capture the minute vibrations. The data was then fed into the customized AI model, which produced transcriptions with around 60 percent accuracy over a vocabulary of up to 10,000 words. While this is far from perfect, the researchers noted that even partial keyword matches could have serious security implications. 'The result was transcriptions of conversations, with an expectation of some errors, which was a marked improvement from our 2022 version, which outputs only a few words,' said co-author Mahanth Gowda, associate professor of computer science and engineering. 'But even picking up partial matches for speech, such as keywords, are useful in a security context.' The team compared their approach to lip reading, which typically captures only 30% to 40% of spoken words but can still help people infer conversations when combined with context. Similarly, the radar-AI system's output, though imperfect, can reveal sensitive information when supplemented with prior knowledge or manual correction. Privacy risks amplified Basak emphasized the potential privacy risks posed by this emerging technology. 'Similar to how lip readers can use limited information to interpret conversations, the output of our model combined with contextual information can allow us to infer parts of a phone conversation from a few meters away,' he said. 'The goal of our work was to explore whether these tools could potentially be used by bad actors to eavesdrop on phone conversations from a distance. Our findings suggest that this is technically feasible under certain conditions, and we hope this raises public awareness so people can be more mindful during sensitive calls.' The U.S. National Science Foundation supported the research, and the team stressed that their experiments are intended to highlight possible vulnerabilities before malicious actors exploit them. They envision future efforts to develop protective measures to secure personal conversations from this kind of remote surveillance. As wireless technology and AI evolve rapidly, this study serves as a crucial warning: even the faintest vibrations from your everyday devices can potentially betray your most private words. The study has been published in, published in the Proceedings of WiSec 2025: 18th ACM Conference on Security and Privacy in Wireless and Mobile Networks. Solve the daily Crossword
Yahoo
06-08-2025
- Yahoo
Lightning Might Be Irradiating Our World Without Ever Showing Its Face. Here's How.
Here's what you'll learn when you read this story: Scientists have long understood why lightning forms, but the atomic processes at the core of the phenomenon have remained largely a mystery—especially the strange mechanics behind terrestrial gamma-ray flashes (TGFs). A new study combines ground-based observation with mathematical models to detail the high-energy photons and X-rays responsible for TGFs. This gives scientists one of the clearest pictures yet of exactly how lightning forms within a thundercloud. When it comes to lightning, it only takes learning a few facts to learn that these brief moments of raw atmospheric power have precisely zero chill. For one, there are roughly 8.6 million lightning strikes on Earth per day, and those terrifying bolts can temporarily superheat the surrounding air to a toasty 50,000 degrees Fahrenheit (a.k.a. five times hotter than the surface of the Sun). Incredibly, lightning can even briefly produce gamma-rays, which are typically spewed from things like supernovae or black hole jets. In other words: when lightning goes, it goes hard. While scientists have a pretty firm grasp on why lightning forms in cumulonimbus clouds, they don't exactly understand the atomic phenomena underpinning the phenomenon. Why, exactly, does lightning produce electromagnetic energy that rivals some of the most intense celestial ongoings in the universe? Now, in a new study from Pennsylvania State University, scientists used mathematical models combined with ground observations to discern the exact atomic workings that create lightning bolts in the first place. The results of the study were published in the journal JGR Atmospheres. 'By simulating conditions with our model that replicated the conditions observed in the field, we offered a complete explanation for the X-rays and radio emissions that are present within thunderclouds,' Victor Pasko, lead author of the study from Penn State, said in a press statement. 'We demonstrated how electrons, accelerated by strong electric fields in thunderclouds, produce X-rays as they collide with air molecules like nitrogen and oxygen, and create an avalanche of electrons that produce high-energy photons that initiate lightning.' Two years ago, Pasko and his team published a Photoelectric Feedback Discharge model, which simulates conditions ripe for lightning activity. And earlier this year, an unrelated study from the University of Osaka observed and detailed the extremely brief moments of terrestrial gamma-ray flashes, or TGFs, using a multi-sensor set-up. They found that TGFs appeared 31 microseconds before connection of the discharge path with the full burst lasting another 20 microseconds after the two discharge paths—one from the ground and one from the air—meet. This new study wanted to understand why these TGFs were often produced without flashes of light or radio wave bursts. Driven by the photoelectric effect, which explains how materials or atoms release electrons when struck by light, the variable strength of the runaway chain reaction that occurs in lightning bolts can precisely produce these initially unintuitive conditions. 'In our modeling, the high-energy X-rays produced by relativistic electron avalanches generate new seed electrons driven by the photoelectric effect in air, rapidly amplifying these avalanches,' Pasko said in a press statement. 'In addition to being produced in very compact volumes, this runaway chain reaction can occur with highly variable strength, often leading to detectable levels of X-rays, while accompanied by very weak optical and radio emissions. This explains why these gamma-ray flashes can emerge from source regions that appear optically dim and radio silent.' Lightning has long been an atmospheric phenomenon on Earth, and one leading theory even speculates that it might have helped kickstart life on the planet. Now, a couple billion years later, that life is finally revealing its high-energy secrets. You Might Also Like The Do's and Don'ts of Using Painter's Tape The Best Portable BBQ Grills for Cooking Anywhere Can a Smart Watch Prolong Your Life? Solve the daily Crossword
