Everyone knows computers and robots only continue to get more and more intelligent as time goes by, and by 2010 bots could already figure out how to bypass any form version of CAPTCHA that has been released so far. This is where reCAPTCHA v2 comes into play.
ReCAPTCHA v2 features the image-based test we have all grown to love(well kinda). Google generated this form of CAPTCHA to be able to teach robots how to identify objects in pictures, you're now probably even more confused and asking the important question: If we are trying to keep robots out, why are we teaching them how to solve it? Well, the answer is relatively simple, we aren't, to put it in a simpler way, we aren't teaching them to solve it rather, to recognize it, and here's how: Have you ever noticed why the CAPTCHAs are always asking the user to identify objects or things in the various pictures that are related to the road and driving? The fact is, Google uses images taken from their satellite and uses you the user to help teach robots to identify things like a fire hydrant or crosswalk in the picture. Google then uses that user data and implement it in their self-driving cars and improve the Google Maps interface and user experience. In this way, the cars while being able to recognize the same things we do and know to avoid them because the last thing we really need is for your driverless Uber to be crashing into a tree on your way to work, have fun explaining that one to your boss.
And yes, you can already guess robots managed to figure this stuff now better than we do, again. In fact, statistics showed for the previous CAPTCHA versions, humans were only able to pass the test with 33% accuracy while robots could breeze through it with a solid 99.8% accuracy. And once again, we had to find a way to beat the system.
No comments:
Post a Comment