Google has killed the CAPTCHA; or at least hidden it from view. The company’s reCAPTCHA is taking a new path for detecting online bots; one that uses machine learning instead of human interaction.
Traditional CAPTCHA operate by getting people to transcribe text that’s difficult for computers to identify. Usually from some grainy picture that are only barely legible. Other systems have tried to get people to match animal sounds to pictures, and listening to numbers being read out while obscured by music.
reCAPTCHA tried to simplify the process by distilling it down to a single button press. It seemed to work, although Google didn’t disclose how it knew that it was a person clicking on the button.
This new version of reCAPTCHA seems to build on the previous mechanics, and has completely eliminated the need for any buttons. In other words, it inserts itself into the process while remaining completely invisible. As before, Google is not saying how it works either.
There’s good reason for Google to remain silent on the inner workings of reCAPTCHA. Even the most basic explanation may provide clues for bot-makers to abuse the system. Considering the number of bots on the internet, this is not some exaggerated threat.
Once the transfer to the new CAPTCHA is done, people should be experiencing a smoother internet experience. Of course, Google is retaining the old CAPTCHA test in the event that it has trouble telling if someone is not a bot.