When we go online, we are being tracked. That's a fact. To a greater or lesser extent, a small cookie or the browser itself sends data to a remote server about the current session. Even the way we type and the words we choose during a search can help form a kind of digital fingerprint about us and our preferences. The Chrome extension called Keyboard Privacy aims to distort the pattern of typing, making it much harder to develop that fingerprint.
A person can identify another by details that seem insignificant from a certain point of view. The scent of perfume, the sound a door makes when it opens, the rhythm of their footsteps... the list goes on. That person transmits these data without realizing it, and something similar can happen in front of the computer. When you visit a webpage and enter your credentials, you develop a certain typing speed, cadence, rhythm. Some keys "chain" faster, and others have a longer delay between them. That is enough to create a digital fingerprint that identifies you, and it is a technique with slow but steady adoption.
At this point, Keyboard Privacy comes into play, an extension created by developer and security consultant Paul Moore. In simple terms, the extension inserts "noise" during our typing, pauses and gaps of about 50 milliseconds on average. The user experience remains intact, and any additional load is briefly transferred to the CPU. We also have the option to disable Keyboard Privacy "per site" in case problems arise with online banking services and similar.
Keyboard Privacy has the ability to defeat profiling systems like KeyTrac (with prior adjustments to the pauses, since 50 milliseconds is a value it can interpret), and although it is currently exclusive to Chrome, its arrival to Firefox is probably a matter of time. Now, what if we use a password manager, or if we decide to "train" to achieve a different typing style? In that case, any attempt to generate a profile will be completely defeated. Personally, I think Keyboard Privacy points in the right direction, but it needs more work. The values of its pauses must be dynamic and random, otherwise any pattern with a 50-millisecond delay will be very easy to record.
