Stanford Security Laboratory’s computer scientists have discovered how to crack audio captchas, using software that can listen in and correctly output the string of random letters and numbers websites use to test whether you’re human, or a malicious bot.
Captchas – short for “Completely Automated Public Turing test to tell Computers and Humans Apart” – are usually images showing a series of characters criscrossed with lines. For the visually impaired, there are audio captchas, which you listen to and have to decode. And now, Stanford researchers have built Decaptcha—a program which recognises unique sound patterns for letters and numbers. After running it through tests, they found that the program works on Digg, eBay, Microsoft, Yahoo, and even reCAPTCHA, a company that creates captchas.
Decaptcha easily beat audio captchas with background noise like static or repetition, but music made its work more difficult. [PhysOrg]