According to a recent study made by a group of researchers from Stanford, when comparing humans to robots – humans don’t do too well in deciphering captcha tests.

The group took more than 300,000 captchas, all are examples taken from popular web companies such as Yahoo! and Google, and gave them to humans (native English speakers) to solve.
When the same captcha was given to three different human subjects, only 71% of the time they had agreed on the right answer.

If to be more specific, Study participants succeeded in solving Google’s and Yahoo!’s CAPTCHAs only 87% of the time, while the ones used by Microsoft were solved correctly only 80% of the time. In audio captcha testing (made for the blind) the results were even worth, as recognition was found to be “as low as 39% for Microsoft and 35% in Google’s case”.

These results support the saying that Captchas – originally designed to keep bots out of all kinds of online platforms and thus prevent spam – are sometime too hard not only for robots but also for humans to decipher.