Scientists from the Republican University of Uruguay demonstrated that with the help of certain equipment and artificial intelligence, it is possible to intercept and decipher the digital video signal displayed on the screen computer and other electronic devices.
In the era of the analog video signal, hackers could easily restore the image on the screen by fixing the leakage of the signal from the video cables. However, with the transition to digital protocols like HDMI, this has become more difficult due to the more complex nature of the transmitted data.
Nevertheless, digital signals still emit a certain amount of electromagnetic waves when transmitting between a computer and a monitor. By teaching an AI model on samples of the initial and intercepted HDMI signals, researchers were able to decode fixed leaks into readable images.
The new technique allows the restoration of text from intercepted HDMI signals with an accuracy of about 70%. While not perfect, this level of accuracy is high enough for the human eye to be able to recognize the text effectively. This potentially enables hackers to monitor passwords, financial data, or encrypted correspondence.
Researchers tested their attack by using software to recognize the text on visual data restored by their AI model. By comparing the extracted text with the original screen contents, they improved the accuracy of recovery by 60% compared to previous methods.
As for the interception of the HDMI signal, hackers can achieve it in several ways. They may install an inconspicuous device to capture a signal inside the building with the target computer or use a radio antenna located nearby to capture HDMI radiation. Real experiments show that electromagnetic leaks can be intercepted even at distances of tens of meters.
Researchers argue that such attacks are already being used against state institutions and industrial facilities. However, these organizations likely protect their premises from electromagnetic leaks despite the significant costs involved.
For ordinary home or office users, this eavesdropping method is still complex to implement. The use of specialized equipment for capturing signals, as well as trained AI models, requires significant costs and efforts.
Thus, data protection resembles a game of cat and mouse. While many modern technologies strengthen digital defense, they also create new vulnerabilities for hacking. Specialists must continuously improve protection systems and anticipate ways they can be bypassed in the ever-evolving digital landscape.