‘There is something uniquely dispiriting about being asked to identify a fire hydrant and struggling to do so’, wrote journalist Josh Dzieza. And who could disagree?
Anyone who browses online will certainly have encountered CAPTCHAs, those security tests designed to distinguish human users from bots and which involve recognising numbers, letters and sounds, or simply clicking to declare “I’m not a robot”.
What does CAPTCHA mean?
CAPTCHA may bring to mind “gotcha”, an informal English term meaning “I caught you”. The association is no coincidence: these tests are designed to catch and stop bots and automated programs.
Of course, it has a more technical meaning:
Completely Automated Public Turing (test to tell) Computers (and) Humans Apart, in other words, the Turing Test.
Devised by Alan Turing, a pioneer of 20th-century cryptography, the Turing Test is a method for assessing a machine’s intelligence by testing its ability to interact during a human-machine conversation. During the conversation, the participant does not know whether they are interacting with an artificial intelligence or another human being. If they are unable to distinguish between them, the test is considered to have been passed.
CAPTCHA and accessibilityFrom letter recognition to Google
Can you hear the Superquark theme tune? We’re about to give you the full explanation.
In the 1990s and early 2000s, the rise in the number of hackers using automated programs on the internet prompted IT experts from various organisations, including those at AltaVista and Sanctum, to join forces to address the new threat. The first CAPTCHAs, based on the optical recognition of distorted letters, provided a defence against automated access to websites.
Users were asked to prove that they were “not a robot”, by correctly identifying and retyping the text displayed.
As CAPTCHA became widely adopted, spam attempts and automated attacks grew more sophisticated. Over time, further improvements were therefore introduced; these included audio-based tests, initially designed to make websites more accessible.
ReCAPTCHA’s “I’m not a robot”
The year 2007 saw the rise of ReCAPTCHA, which originated from a research project at Carnegie Mellon University and transformed the very concept of CAPTCHA.
The basic idea was to create a system that would protect websites from spammers and bots while also helping to digitise books and texts that computers could not read. In other words, by completing the CAPTCHA, users were supporting the digitisation of books and documents and making them accessible online.
After acquiring the company in 2009, Google launched the No CAPTCHA ReCAPTCHA” system in 2014, which was particularly well received for its ease of use: a simple click or tap is enough to prove that “I’m not a robot”.
Using advanced technologies such as machine learning and user behaviour analysis, the US technology giant’s system can accurately determine whether or not the user is human.
How does a CAPTCHA work?
It depends on the type.
- I’m not a robot: the parameter used for the Turing test is based on the user’s overall behaviour before and after clicking. It therefore includes mouse movements, subsequent actions, previously visited websites and other similar variables.
- Distorted text: users are asked to decipher a distorted image by converting it into characters and demonstrate their ability to interpret complex visual inputs. Variants may include rotated, distorted or overlapping letters, together with a background that makes interpretation even more difficult. An audio version is also available.
- Images: while we humans can easily distinguish objects in a photo, such as cars or traffic lights, this is more difficult for a computer. It must interpret the request, examine the image and identify where the objects are located. This makes image-based tests more secure, but also adds an extra level of complexity compared with text-only tests.
The drawbacks of CAPTCHAs
Despite its many advantages, CAPTCHA technology also has some drawbacks.
User experience: online procedures become longer, causing inconvenience for some users. The increasing complexity of these tests makes completing them a frustrating experience, as highlighted by a study conducted at Stanford University in 2010.
Low conversion rates: a poor user experience and inaccessible tests can cause a decline in website conversions. According to a 2009 study involving 50 websites, the use of CAPTCHA caused a 3.2% decrease in legitimate conversions.
Privacy concerns: some users and researchers are critical of AI-driven CAPTCHAs such as reCAPTCHA v3 because they use code and cookies to track users’ activities across different websites. This raises questions about transparency and how the data collected is used.
Unusual ways to say ‘I’m not a robot’
The web is full of strange CAPTCHAs. Some users have been asked to pick out the happiest-looking dogs from a series of images, while others have had to identify clouds that look like horses. A simple task? If it were not for the elephant-shaped clouds.
There are also CAPTCHAs that test you with graffiti in different artistic styles, asking you to pick out the letters of the code among the drawings, or to identify a geometric shape that seems to defy the laws of physics.
And if you are wondering whether there are any more unusual ones, the answer is yes. In 2014, Kevin Gimpel and his team created a CAPTCHA that compares photos of muffins and seal pups. The task is to select only a specific category of images from a series of deceptive and distorted representations.
Good luck!
