The publication of a joint study conducted by researchers from the Swiss Federal Institute of Technology in Zurich, the University of California at Irvine and Microsoft is likely to question the use of CAPTCHAs as a means of proving that visitors to a website are indeed human beings and not robots.
In the preamble to their scientific publication, Professors Searles, Nakatsuka, Ozturk, Paverd, Tsudik, and Enkoji declare: “For almost twenty years, CAPTCHAs have been widely used as a means of protection against robots. Over the years, as their use has grown, techniques for defeating or circumventing CAPTCHAs have steadily improved “.
CAPTCHAs are programs used on many websites to prevent credit card fraud, for example, while cybercriminals are increasingly using automated programs to hack into systems. Some challenges consist, for example, of choosing the images representing traffic lights or cars. In essence, they are visual tests that only human beings can complete because they are complicated.
Bots have become stronger than humans at solving CAPTCHAs
But that’s for theory, because in practice advances in machine vision and artificial intelligence now allow machines and bots to respond as well, but more importantly, faster than human subjects. As part of this study, the scientists asked each of the 1400 attendees to solve and evaluate the level of difficulty of ten captchas submitted by 120 of the most popular websites in the world.
For the scientists, the conclusion of their research is clear: robots finish tests faster and more accurately that we. Indeed, with CAPTCHAs, web developers and cybercriminals engage in a veritable game of cat and mouse. But the consequence for Internet users is that the tests to be solved are always more complex and tedious. Is it time to abandon this authentication technique? The answer is yes for Google and Apple, who want the skin of CAPTCHAs and offer an alternative.