Hong Kong researchers have identified a new type of eavesdropping attack that can threaten the privacy of users of headphones, phones and some smart home devices. This method, called injective, unlike traditional electromagnetic eavesdropping attacks, instead of passively receiving weak signals, first actively injects a radio signal into the device.
According to ITNA quoting Techspot, this research was conducted by researchers at the Hong Kong University of Science and Technology in Guangzhou and The Hong Kong Polytechnic University, and its paper was presented at the USENIX Security 2026 conference. In their experiments the researchers examined 11 commercial products, among them headphones from Sony, Apple, HP and Philips, Ugreen headphones, a VoIP phone and a number of fans and smart lights from various brands.
The basis of the attack is the presence of nonlinear components in the electronic circuits of devices. The attacker sends a radio signal to the device in the range of zero to 9 MHz. This signal can interfere with components such as amplifiers, analog-to-digital converters, power converters and switching transistors. As a result, part of the audio information or device activity is emitted in the form of a receivable signal.
Using specialized radio equipment, the researchers were able in many experiments to receive information from a distance of more than two meters and even from behind a wall. The typical attack range for different devices was between one and six meters, but using a radio frequency amplifier, intelligible headphone audio was recovered from distances up to 30 meters.
One important point of this attack is that encryption, randomization and other software protections cannot directly stop it; because the information leakage occurs on the hardware and analog path, not in the encrypted digital data.
This attack still requires proximity to radio equipment and information about how the target device will respond. However, experiments in environments such as hotel rooms and behind furniture have shown that the attack idea is not merely limited to laboratory conditions. The researchers suggest that using electromagnetic shielding, filters and twisted-pair wiring can reduce the likelihood of success of such attacks, although there is no guarantee of complete prevention.