Captchas are the very annoying things where you have to type a set of letters and numbers in order to prove that you are human rather than robot. Why are they necessary? Because so many things have become automated and some companies don’t want to provide service to robots or waste their resources on fake requests. Similar to Turing tests, but still very different.
Existing captchas are generally in the form of text presented as images. (There are other methods but this seems to be the most popular one at the moment.) This in itself presents a heck of a lot of problems. The biggest problem is that they’re not accessible to people who can’t see and depend on screen readers. The workaround has been to provide an audio version of the text but have you tried listening to the sound clips? They’re read so fast and there’s so much background noise that I don’t know how you can figure out what they’re saying. I suppose that if you’re used to using screen readers (which can read at super high speeds), you may have less problems hearing the text.
Another big problem with the captchas is that even if you can see the image, you’re not always sure what the text says. There are algorithms in place that distort the text so that a computer can’t use image recognition on it to detect words. In attempting to fool a computer, it’s also fooling a human being. It can get quite frustrating and time consuming to type the words numerous times just to complete what should have been an easy task (like logging in).
Technology has become so advanced that it is no longer about trying to tell if you are human or a robot, but rather that we have to prove that we are human. The burden has been put on us to prove that we’re not a robot. The existing solutions are a good attempt at keeping away from robots but they’re also keeping away humans. I know that there is research happening that is trying to resolve the problems such that it would be easy for humans but difficult for robots (Captcha.net and this paper from Towson University and University of Notre Dame have some examples). It is an extremely difficult problem and may take a long time to completely solve.
Recently, I had gone to Hong Kong and while there, Facebook required me to prove that I am me since I was logging in from an “unusual” location. It showed me a picture of one of my friends on my Facebook friends list and asked me who was in the picture. This process was repeated numerous times until the system decided that I really was the person I said I was. I thought this was a really good way of proving our identity. Yes, it would be frustrating to go through this every time we logged in but I feel it is heading towards a good direction. It’s a better solution than those security questions which we don’t always remember the answers of (even when we created them ourselves). It requires access to your personal information, which you don’t want all websites to have access to. The solution is not perfect and does not work for all situations, but it certainly is unique.
