“Forgot password?” This phrase is haunting to most of us, especially with the plethora of accounts we maintain, so forget about forgetting a password. It is mainly about remembering and maintaining the passwords, pins, OTPs, and many other keys that need our input every time we need to access our accounts. The other alternative is biometrics which works well, but it’s too invasive for some users and it gives them a sense of discomfort.
A niche category in the field of biometrics is Keystroke Dynamics. Çeker et al. wrote a paper on a study titled “User Authentication with Keystroke Dynamics in Long-Text” and was presented at the 2016 IEEE 8th International Conference on Biometrics Theory, Applications, and Systems (BTAS). This study was conducted on 34 people in two sessions in a lab at Clarkson University over a period of eleven months. The result was a hundred percent accuracy with AUC (Area Under the Curve) = 1.
Çeker et al. incorporated the methodologies of other studies done on Keystroke Dynamics and found out that this form will give 100% accuracy if the user verification is done through active authentication. This would run software that continuously collects the typing patterns in the background and with the usage of fewer diagraphs. The researchers said that “For an active authentication mechanism, a system should authenticate a user unobtrusively and transparently in a continuous manner.”
https://twitter.com/TypingDNA/status/1356964608309948419
Understanding typing patterns
Every individual has a unique DNA, and its uniqueness is represented physically through fingerprints, retina, or the extraction of DNA from blood and bodily fluids. There are other representations of our DNA as well, such as our behavior. And one of such behaviors is our typing behavior. Keystroke dynamics is a form of behavioral biometrics that provides continuous authentication of users while working at a terminal.
Individual typing patterns recognition systems analyze the way a user types at a terminal by monitoring the keyboard events. “In other words, not what is typed, but how it is typed is important,” writes Chora et al. in their paper titled “Recognizing Individual Typing Patterns“. They recognized some keyboard events can be analyzed, such as time between key-pressed and key-released events, break between two different keystrokes, the duration for digraphs and trigraphs, and others.
Diagraphs are two keys typed one after the other, while trigraphs make a three-keys typing event. The duration of digraphs or trigraphs is measured between a pressing event of the first key and the release of the last key. Combining these several keyboard events to recognize typing patterns will provide uniqueness to an individual to manage their verification.
Biometrics like Thumb prints & Face ID are popular but have you seen the mind blowing tech from @TypingDNA?!
The way you type is your authentication!
Check out the demo! Quickly test it by yourself or with a friend & see how it works!
What do you use for authentication?#ad pic.twitter.com/vcpshi8T5T
— Danny Thompson (@DThompsonDev) March 11, 2021
How TypingDNA works
TypingDNA is a New York-based company that provides user authentication through the Keystroke Dynamics method and has just launched its services to the market last March 16. The main focus of the company is to replace the age-old techniques of user verification, such as SMS 2FA codes, and minimizing user effort.
At the moment, they are providing free services for the first year on a first-come-first-serve basis to companies. To improve the UX and bypassing the route to the “no-app-needed” method, TypingDNA wants to use its tools to make the typing patterns one of the unique signatures of a human being.
TypingDNA Verify uses its AI-powered proprietary typing biometrics authentication technology to select the best keyphrases enabling higher matching accuracy or shorter texts. Their biometrics engine uses RESTful Authentication API to analyze and match typing patterns with a known or an enrolled user. The recorded typing pattern can be sent to their API via RESTful request store or match against a previously recorded pattern. Then the API will output a JASON response to the request, which will contain the result of authentication.
The Typingdna.js is a function that calls the website to match a requested typing pattern. The ‘TypingDNA JavaScript recorder class’ is implemented on the front-end of their web app. All the requests to the API should be made from the application’s backend. The TypingDNA class is open source, available under Apache Licence (Version 2.0). The source code is public and can be downloaded from the GitHub repository.
We’re thrilled to announce TypingDNA Verify! Check us out on Product Hunt today: https://t.co/VjJ62Uae4q
— TypingDNA (@TypingDNA) March 16, 2021
Conclusion
“According to recent research, 73 percent of consumers say that the process should happen instantaneously when trying to create an account. Furthermore, an overwhelming majority (92 percent) expect a fast, frictionless experience,” writes Silvana Chirita on the company’s blog. “The smart, user-friendly authentication” will be available for just 1 cent per active user per month, regardless of the number of typing biometrics verifications. TyingDNA Verify is designed to support companies dealing with a large number of authentications a month. “The first companies to integrate the service will get access free of charge for the first year for up to 100,000 end users,” writes Chirita.
Currently, the company is also conducting public research on recording various typing patterns all over the world. You can also be a part of this research by going to the company’s website. Joint Research with TypingDNA and also get your mood analyzed. Say, if you are feeling stressed and are going to participate in the research, you can preset your mood in the MoodSettings, and the typing samples will be collected centered around how the user is feeling.
YouTube: TypingDNA Verify: Smart, user-friendly authentication replacing SMS 2FA codes
Photo credit: The feature image is symbolic and has been taken by Juan Gomez.
Sources: Hayreddin Çeker and Shambhu Upadhyaya (IEEE Xplore) / IEEE Xplore / Michał Choraś and Piotr Mroczkowski (SpringerLink)
