All of us have, at some time or another, found ourselves squinting unattractively at our computer screens, desperately trying to decipher the mangled letters or numbers presented by a CAPTCHA. CAPTCHA actually stands for Completely Automated Public Turing Test to tell Computers and Humans Apart, and the frustrating little puzzles have been a standard part of web protocol for years. By presenting a visual recognition challenge that computers and bots find difficult, they have thus far been a means to prevent spamming and other automated attacks against websites – but it’s all about to change.
Up until this point, CAPTCHAs have consisted of a single word that a computer can recognise, and one that it can’t: the job of you, the human, is to identify both correctly. By using distorted words from archives and historical documents, CAPTCHAs have also been used to digitise massive amounts of literature, which is all very commendable, if still infuriatingly difficult. Google’s studies, however, report that over 90% of people find CAPTCHA tests difficult, and even more worryingly that computers can now solve an estimated 99.8% of them. So they’ve got to go.
Traditional CAPTCHAs are now being replaced with Google’s new ‘No CAPTCHA reCAPTCHA’, which involves ticking a box. Yes, it’s that simple: tick this box which says ‘I’m not a robot’, and you’re done. How does this prove you’re human? Because you’ll do it wrong. Well, not wrong, but imperfectly. The minute, imprecise movement of a mouse controlled by a human hand, as well as unpredictable hesitations are all representative of your cumulative humanity, and difficult for computers to replicate.
The change in protocol, however, is worth some consideration. It has become necessary due to the increasing complexity of computer analysis, and it’s worth questioning how long the new CAPTCHA tests will be viable. The emergence of ‘smart computing’, what we may one day point to as the precursor of AI, has been and will continue to be one of the game-changing developments in computing this century.
Whilst a relatively small development in the grand scheme of things, the end of traditional CAPTCHA as a means of flummoxing computers may be the beginnings of a trend. Stephen Hawking’s highly publicised warnings about AI might be a little premature, but it is entirely possible that computing and industry standards that have heretofore stood the test of time will be undermined by a new generation of reactive, smart computing. It all sounds very science fiction, but as our computers evolve so must our processes, to ensure both our own security and, of course, to prevent the rise of the machines.
Image courtesy of Becky Stern