In an increasingly interconnected digital world, the Robot threatens Malicious has evolved drastically. These automated programs, designed to simulate human behavior, They represent up to 50 % of Internet traffic And they are used in advertising fraud, in the creation of false accounts and attacks on biometric systems.
Its detection implies the design of an additional level of security that protects sensitive data. However, current systems, such as codes CaptchasHim Traffic analysis and the Multifactive authenticationThey have significant vulnerabilities.
The most advanced robots manage to overcome these barriers, exploiting artificial intelligence techniques to copy human behaviors almost perfectly.
The role of behavioral biometries
The robots have reached an alarming level of refinement: The mimic the movements of the mouse, the cadence of the click or the writing on touch screen. This raises a crucial challenge for traditional security systems that are often unable to distinguish between a true human being and a bot.
Given this reality, it is essential to improve and evolve detection methods, developing solutions that can adapt to emerging threats.
This is where the Behavioral biometriesTechnology that identifies users through their behavior models by interacting with digital devices. This is a new line of defense that promises to change the rules of the game, in particular the analysis of the stamp of the human movement.
Like the Timbre distinguishes a human voice of a synthetic, in the human movement, the The stamp reflects unique irregularities derived from biomechanical and neuromotor processes. These characteristics are extremely difficult to emulate for robots, which makes them a distinctive feature for detection.
Unmask robot that imitate humans
In this scenario, at the University of Las Palmas de Gran Canaria we developed the Biotimbre projectwhich proposes a disruptive approach based on the Saed cycle (acronym for synthesization, attack, evaluation, detection).
This not only tries to strengthen robot detection, but also to improve their simulation ability through the advanced synthesis of human movements. Therefore, the first step is to summarize: the most realistic robots are created by incorporating the “bell” of the human movement, which includes the natural imperfections of neuromotor and muscle systems.
These generated robots are used to simulate attacks and evaluate the effectiveness of current detectors. The results obtained are analyzed to identify weak points in detection systems.
Thanks to this, the models entertainly improve through training with increasingly realistic samples.
Synergy between generation and detection
The innovation of the Saed cycle lies in its ability to improve both ends of the IT security spectrum. On the one hand, we create more powerful robots: including the bell in generative models, robots become indistinguishable from real human movements.
On the other hand, more effective detectors can be designed: exposure to more advanced robots allows you to form detectors that exceed the skills of current systems and increase their precision and robustness.
Future implications
The analysis and synthesis of the bell do not only have applications in IT security. Its potential extends to sectors such as health, to monitor movements in patients with neurodegenerative diseases and education, to evaluate motor development in children, for example. This approach offers a new standard in human-macchine interaction and the protection of our digital identity.
The fight against robots is a constant challenge, but using our unique behavioral shots, we can be a step forward in the safety tender.
Moisés Díaz CabreraProfessor of applied physics, University of Las Palmas de Gran Canaria y Miguel Ángel Ferrer BallesterProfessor of sign theory and communication, University of Las Palmas de Gran Canaria
This entry was originally published in The conversation. Read the original.
