Google’s bot-detecting technology reCAPTCHA is going invisible.
The system, used by millions of websites across the Web, is used to recognise human users and filter out automated web crawlers; its interface is familiar to whoever has surfed the Internet in the last decade: it can variably ask users to make sense of slightly warped words or numbers, to select pictures showing specific images, or to tick a self-exculpatory checkbox reading “I’m not a robot.”
reCAPTCHA started off as an autonomous company, before being bought off by Google in 2009. Since then, the tech giant has managed to leverage it to kill two birds with a stone: when users identified words or images to prove their humanity, they were also unwittingly aiding Google in training its image-recognition algorithms.
Now, though, Google has announced reCAPTCHA queries will no longer bother web users. The new version will work in the background and use “a combination of machine learning and advanced risk analysis” to pick up subtle cues in order to distinguish humans from bots.
The Mountain View-based company has not revealed the new system’s exact workings, presumably to prevent bot-makers from, designing new workarounds.
What is known, though, is that, were a user to fail the “invisible” reCAPTCHA test, they will be asked to solve the good old image recognition or checkbox-ticking challenges.
Image via bit.ly/2mIIfbH
Co-working space and blog dedicated to all things data science.