ARTFEED — Contemporary Art Intelligence

Smartphone Touchscreens Leak Handwriting via Electromagnetic Side Channel

ai-technology · 2026-05-16

A significant security flaw has been identified in capacitive touchscreens, as researchers have shown that electromagnetic (EM) emissions produced while writing on smartphones can be intercepted and reconstructed into identifiable two-dimensional paths. The attack method, referred to as TESLA (Touchscreen Electromagnetic Side-channel Leakage Attack), employs a non-contact technique to capture EM signals from on-screen writing, converting them into real-time continuous handwriting. In experiments conducted on various commercial off-the-shelf (COTS) smartphones, TESLA achieved a character recognition accuracy of 77% and a Jaccard index of 0.74, reflecting a strong resemblance to the original handwriting. These results were published in the Cryptography and Security section of arXiv (ID 2512.11484), revealing a new side-channel vulnerability for touchscreen devices.

Key facts

  • TESLA attack recovers handwriting trajectories from EM emanations of capacitive touchscreens.
  • Non-contact framework captures EM signals during on-screen writing.
  • Achieves 77% character recognition accuracy on COTS smartphones.
  • Jaccard index of 0.74 for trajectory similarity.
  • Published on arXiv under Computer Science > Cryptography and Security.
  • Paper ID: 2512.11484.
  • Exploits electromagnetic side channel of touchscreens.
  • Real-time reconstruction of 2D handwriting trajectories.

Entities

Institutions

  • arXiv

Sources