Bugged with passwords? Your daily activity could help you log-in
| Tuesday, November 3, 2015 - 11:09
Kolkata: Bugged with multiple passwords and log-in details? Lighten up. Thanks to a new technology, data extracted from your daily activity, such as those on Facebook or phone, could be used for authentication.
A team of researchers belonging to the Indian Institute of Technology-Kharagpur (IIT-KGP), the University of Texas at Austin, US and the University of Illinois Urbana-Champaign, US may have hit upon an alternative system that will allow you to bypass password-based authentication on your personal devices.
Major benefits of the 'ActivPass' system include freedom from remembering multiple complicated passwords and thwarting the practice of password-sharing for cloud-based services such as Netflix and Hulu.
In a paper titled "ActivPass: Your Daily Activity is Your Password", the team shows how it is possible to extract "adequate secrets" by observing the user's activity logs from Facebook, browsing history, call logs, and SMSs and then use those to frame questions on infrequent and unpredictable happenings.
For example, in order to access a certain website on your smartphone, you could be asked "Who called you from Mumbai last evening?" or "Which song did you listen to during lunch hour today?"
"Our detailed experiments reveal that people can remember their recent activities with a fairly high degree of ease and when the chosen activities are outliers, it is much too difficult for impostors to answer questions pertaining to those.
"Thus ActivPass can help rid us of the necessity to remember and periodically change the credentials for the hundreds of accounts that a user maintains today," said Niloy Ganguly , co-researcher from IIT Kharagpur, a key member of the research team, in a statement on Monday.
With security improvements in the future, activity-based authentication could fill in for the inadequacies in today's password-based systems, according to the paper.
For the project experiments involving 70 volunteers recruited from various segments of university population, it was found that in 95 percent of cases, the system successfully authenticated a legitimate user.
The team has already developed a promising primary prototype and is presently working on improving the system with addition of personalization features and privacy-preserving architecture to be incorporated in a full-fledged version that may be commercialized.