Surprising research! Can a fiber cable turn into a listening device?

Fiber optic internet cables have been one of the most robust parts of secure communication infrastructure for years. The fact that they are not affected by electromagnetic interference, do not emit radio frequency, and can carry data with low loss over long distances strengthened this perception. However, a new study presented at NDSS 2026 revealed that these cables can become a side channel for acoustic listening under the right conditions.

The study prepared by researchers from Hong Kong Polytechnic University and the Chinese University of Hong Kong shows that FTTH fiber connections coming to the home and office do not only carry data, but can also carry traces of sound-induced micro vibrations in the environment. When these traces are measured with special equipment, data about human movements, in-environment location information and some speech contents can be obtained.

How does acoustic listening with fiber optic cables work? At the center of the research is the sensitive structure of optical fibers against mechanical vibrations. As sound waves travel through the air, they can cause very small deformations in the surrounding surfaces and fiber cables. These micro-changes in the structure of the fiber create measurable differences in the backscatter and phase of the light passing through the cable.

Distributed Acoustic Sensing, or DAS system, can read these differences and extract the vibration information that the cable is exposed to. At this point, the attack works differently than a classical listening device. There is no need to have an electronic insect in the environment that carries a battery, emits a radio signal, or acts like a microphone. By monitoring optical changes at one end of the fiber line, the attacker can analyze the vibrations at the other end.

Therefore, traditional RF scanners or classical bug detection methods may be insufficient to directly detect such an attack. However, there is an important distinction here. A standard straight fiber cable running through a wall is not sensitive enough to clearly pick up human speech on its own. In the researchers’ tests, when a speaker playing speech at 80 dB was 1 meter away from the fiber, speech could not be recovered.

On the other hand, footsteps at a level of approximately 76 dB can be detected. In other words, we cannot say that the existing fiber cable in your home can listen to you without any additional elements. The main element that makes the study remarkable is the passive mechanism that the researchers call Sensory Receptor. This structure is created by wrapping standard telecom fiber around a hollow cylinder. Sound pressure first creates a small deformation in the cylinder, this deformation is transferred to the fiber as mechanical stress, then the length and phase changes in the fiber turn into a measurable sound signal.

When viewed from the outside, this mechanism can be compared to an ordinary fiber box. This increases the importance of the attack in terms of physical security. Because for the attack to be strong, there must be a receiver mechanism close to the environment on the victim’s side. This mechanism is not an actively broadcasting electronic bug, but it is still a special structure that is physically placed or hidden inside the fiber box.

In some posts on this subject on social media, we can see statements such as listening can be done from 50 meters away. What is meant here is not that the conversation in the room was captured from the air from 50 meters away. In the scenario, DAS equipment at one end of the fiber line can read the vibrations occurring in the fiber mechanism located close to the target environment at the other end of the line. In other words, the measurement equipment may be far away, but the fiber section affected by the sound must be close to the target environment.

Moreover, the research does not only focus on the speech listening side. Different scenarios are also being tested, such as classifying sound events, estimating the location of the sound source indoors, and retrieving speech content. It talks about 14 different sound events, indoor positioning with three receivers, and post-noise reduction reprocessing of speech samples. On the speech side, the results are remarkable, but not perfect.

Distance, ambient noise, location of the fiber, structure of the receiver and angle of the sound source directly affect the result. While more understandable data is obtained at close distances, the word error rate increases as the distance increases. Therefore, the study does not show that each fiber line turns into a high-quality microphone, but that a new side channel is formed that affects speech privacy in special conditions.

Another important part of the research is ultrasonic jamming tests. Ultrasonic jammers used to jam traditional microphones do not silence fiber-based detection in the same way. This shows that fiber-based attacks create a different threat model than classical voice recorders. Methods such as microphone searching, RF scanning or ultrasonic jamming alone do not provide sufficient protection in this scenario. This risk becomes most significant in areas where sensitive meetings are held.

Corporate offices, law offices, financial institutions, public buildings, embassies, defense industry companies and R&D centers are in a more critical position under this heading. For home users, the picture is far from panic level today. For the attack, access to the correct fiber line, commercial DAS equipment, passive receiver mechanism close to the target environment and advanced signal processing are required. The first step on the defense side is to treat the fiber infrastructure not only as a network connection but also as a physical security element.

In sensitive rooms, unnecessary fiber excess should not be left, cables should not be wrapped in a loop, they should not come into contact with surfaces that will increase vibration, and fiber boxes should be checked regularly. Researchers also consider optical isolators and connection solutions that could disrupt DAS measurement as possible countermeasures.