If you think that after locking the smartphone, turning on flight mode or turning off the internet, your privacy becomes completely safe, then you are wrong. Because a new research from China may surprise you. Researchers at China's People's Public Security University have claimed that they have developed a technology that can identify which app is running on the phone and what the user is doing by analyzing the very weak electromagnetic radio signals emanating from the smartphone. Researchers say that for this technology there is no need to unlock the phone, access the operating system or extract the data present in it. This research has been published in the peer review journal Radio Engineering on 22 May. However, at present it has been tested only in a controlled lab environment.
Researchers at People's Public Security University of China have developed this technology for digital forensics. According to the research paper, this is a non-contact forensic technique, that is, information can be collected without touching the phone or accessing its operating system and stored data. Researchers say that the smartphone emits very weak low frequency electromagnetic signals while working. These signals keep changing during different apps and user activities. By analyzing these signals, it can be found out which app is being used in the phone. According to research, this method can help in gathering additional technical evidence during digital investigation.
The research team tested this technology on Apple iPhone 15 Pro, Xiaomi 15 Pro and Oppo Reno 13. According to the researchers, the model identified which mobile app was being used on the phone with 99.07 percent accuracy. The test included apps like Douyin, WeChat video call, Baidu Maps, SMS, browser, camera and cloud storage. Not only this, this model also identified with 95.61 percent accuracy whether the user is pausing the video or audio, playing it at normal speed or watching it at double the speed. This testing was done on platforms like YouTube, Bilibili and Tencent Video.
According to the research paper, every mobile app uses the phone's hardware in a different way. Some apps put more pressure on the processor and graphics chip, while others activate the GPS, Wi-Fi module, storage controller or wireless communication circuits. This changes the power consumption pattern and produces different types of low frequency electromagnetic signals. For example, graphics processors and decoding hardware are more active during video streaming, while navigation apps use GPS and storage from time to time. By recognizing these different signals, Artificial Intelligence guesses which app and what type of activity is going on in the phone.
Researchers placed a special induction coil on the back of the smartphone, which was connected to a digital monitoring receiver. This instrument recorded electromagnetic signals between 150 kHz and 650 kHz. After this, these signals were converted into time-frequency spectrograms and analyzed with the help of Spectrogram Transformer deep learning model. Researchers say that this method can help investigative agencies in gathering technical signals related to behavior even without extracting phone data. He has described it as a useful technique to investigate in special circumstances without attracting attention.
Although the researchers have claimed the accuracy of this technology, they have also admitted that all the tests were conducted in a controlled lab environment. In this, the high precision digital monitoring receiver was kept very close to the smartphone. The research paper did not say how effective this technology would be in crowded electromagnetic environments, long distances or in real conditions. The team says that in the future, work will be done on identifying previously unknown devices, creating a separate electromagnetic fingerprint of each device and collecting signals from less specialized devices. Researchers have also suggested that in future, the possibility of collecting such signals will also be explored by using magnetometers of smartphones or electrodes of wearable electronic devices.
