Alex Hern 

Don’t click on the traffic lights: upstart competitor challenges Google’s anti-bot tool

New charges for reCaptcha spur web security firm Cloudflare to seek out an alternative
  
  

Google logo and world map
Google’s test to check identity, reCaptcha, has dominated worldwide since 2009, but the company is now charging some users. Photograph: Andre M Chang/Rex/Shutterstock

The days of clicking on traffic lights to prove you are not a robot could be ending after Google’s decision to charge for the tool prompted one of the web’s biggest infrastructure firms to ditch it for a competitor.

“Captcha” – an awkward acronym for “completely automated public Turing test to tell computers and humans apart” – is used by sites to fight automated abuses of their services. For years, Google’s version of the test, branded reCaptcha, has dominated, after it acquired the company that developed it in 2009 and offered the technology for free worldwide.

Google’s introduction of charges for the service has prompted Cloudflare, a little-known firm that protects around 12% of the internet from bot attacks, to seek an alternative.

The company’s founder and chief executive, Matthew Prince, said: “It would have added millions of dollars in annual costs just to continue to use reCaptcha for our free users. That was finally enough of an impetus for us to look for a better alternative.”

Cloudflare had long been uneasy about using Google’s service, Prince added. “Some customers have expressed concerns. Google’s business is targeting users with advertising; Cloudflare’s is not. We have strict privacy commitments. We were able to get comfortable with the privacy policy around reCaptcha, but understood why some of our customers were concerned about feeding more data to Google.”

China’s block on Google was an additional concern, he said.

Cloudflare will instead use a similar service, hCaptcha, developed by Intuition Machines, a California startup. The tool has a wider range of tests than reCaptcha, including one asking users to draw a box around a specific object, or to identify the correct label for a picture.

Google’s reCaptcha had been free for so long because, as well as testing whether a site visitor is a robot, it uses humans to do menial tasks. The first version asked users to transcribe hard-to-read words from scanned books, while later versions used similar technology to get people to label images. This helped Google’s image recognition software to tell the difference between, say, cats and dogs, and later taught the company’s self-driving cars and mapping vehicles how to recognise street signage.

The tool was developed in part by the Guatemalan entrepreneurLuis von Ahn, who used a similar idea to create the first version of the language-learning app Duolingo.

In the past, that value exchange was enough to keep Google’s tools free, but recently the company added an enterprise tier for corporate users who make more than a million calls a month, charging undisclosed fees. “In our case,” says Prince, “that would have added millions of dollars in annual costs just to continue to use reCaptcha for our free users.”

Users might not immediately welcome the replacement, however, because for all the difference behind the scenes,the tests that hCaptcha poses can be very similar to Google’s. A sample on hCaptcha’s website asks users to click through 9 to 27 images and select those which show a train.


• This article was amended on 17 April 2020 because, due to an editing error, an earlier version said hCaptcha was created by Cloudflare.

 

Leave a Comment

Required fields are marked *

*

*