Google Kills CAPTCHA With Auto Human Detection Tests

03/12/2014 14:42 GMT | Updated 03/12/2014 14:59 GMT

Google has killed CAPTCHA.

The online security widget, through which users 'prove' they are human by typing out nearly-unreadable passwords, is nearly ubiquitous online -- despite the fact it barely works.

It started with great hopes. The "Completely Automated Public Turing test to tell Computers and Humans Apart" as it's properly known was initially very successful.

But the internet has a way of corrupting beautiful things, and making them useless. And so it proved with CAPTCHA. For while managed to weed out a fair bit of spam at first, its effectiveness dropped markedly. In fact, Google recently showed it could crack CAPTCHA with a machine 99.8% of the time.

Fact is, it also bamboozles users. Text in CAPTCHA tests is by this point almost always garbled to the point of nonsense, the photos are low-resolution and hard to read and the end result is - for many - giving up altogether on trying to access whatever it was you were doing.

Well Google has a new system. It's called No-CAPTCHA, and the idea is to use more data about your habits and behaviour to tell if you're human or not.

"We figured it would be easier to just directly ask our users whether or not they are robots," it said. "So, we did!

Using the new API, Google will pre-screen users and filter through those who are obviously human. If that's you - great, you'll just see a checkbox and move on. Google hasn't said exactly what that data is, but it won't be collecting anything about you it doesn't have already.

But if it's not sure - then you'll have to pass another test. Though you won't necessarily be typing anything. More likely you'll be sorting images, or doing something else vaguely useful for Google's search algorithm but also friendly for touchscreens.


One example has users search through images of dogs and cats, picking out only the cats.

It sounds great, and like it could be a big upgrade for the web. Google said: "Humans, we'll continue our work to keep the Internet safe and easy to use. Abusive bots and scripts, it’ll only get worse—sorry we’re (still) not sorry."

And we're sure there's nothing terrifying about Google owning an automatic human detection machine. Nothing at all. Not with all its robots. Er...