Keystroke Dynamics and the future of personal data

"Perhaps we have not seen a great diffusion of Keystroke Dynamics in our day to day, compared to other access methods such as the famous two-factor authentication (2FA), but there are companies in the market like TypingDNA that have benefited from the development of this technology."
Por Genaro Almaraz
Nov 23, 2021

All the people are different. We have physiological and biological traits that characterize us and these have been used in the field of cybersecurity as methods of access to our mobile applications and personal data. These traits range from our retina and DNA, to even the way we type on the keyboard. Keystroke Dynamics is the name of the field of study where the entity under observation is the writing patterns of the users.

Keystroke Dynamics may seem like a new topic to us, but in reality it has been in scientific knowledge for years. For example, a study developed in 2002 by the Center for Artificial Intelligence of the Tecnológico de Monterrey and its department in Computational Sciences, made software to compare statistical methods and Data Mining as approaches for identifying users through keystrokes. keyboard (Francisco J. Gutiérrez et al., MICAI 2002).

A question may arise here: exactly what are the characteristics that define unique patterns of writing on a keyboard? Well, we can list the following:

  1. The duration of a key press.
  2. How often do we make mistakes when writing.
  3. How long does it take us to go from one key to another.

Specifically in the Tecnológico de Monterrey article mentioned above, the time it took to press one key and switch to another were the two most relevant factors for their analysis: How long it takes us to go from one key to another.

If someone is interested in delving deeply into how to create these classifiers, I recommend reading techniques such as Bayesian Networks, Clustering and Decision Trees, because in general, compared to statistical methods such as Regression or simply the calculation of means and standard deviations, Artificial Intelligence methods allow a better control of the hyper parameters of the model from its tuning.

Perhaps we have not seen a great diffusion of Keystroke Dynamics in our day to day, compared to other access methods such as the famous two-factor authentication (2FA), but there are companies in the market such as TypingDNA that have benefited from the development of this technology.

As potential use cases of Keystroke Dynamics we can mention the following:

  1. The signing of PDF documents.
  2. Preventing cheating on an exam.
  3. The development of virtual assistants for the generation of reliable medical reports.

Thanks to science, technology and business development come, and this is no exception for advances in computer security.

With all this I am not saying that Keystroke Dynamics will break all the limits of the market, but it can be a great complement to the password-based methods that we currently live with on a day-to-day basis.

Computer security is a responsibility that we have to assume strongly, because, without a doubt, as human beings we are the weakest link in the protection of information.

Genaro Almaraz

Genaro Almaraz

Master in computer science.
Share This