Machine learning algorithms are so powerful that even humans cannot pass Google CAPTCHA

Original article: The Verge
Sina Technology compiled and sorted

If you’ve found that the various authentication codes used online are becoming increasingly difficult to recognize, you’re definitely not alone. This is all because artificial intelligence is getting smarter.

This type of verification code is called CAPTCHA (Completely Automated Turing test to tell Computers and Humans Apart) in the industry, and The Verge's latest report reviewed the history of CAPTCHA and found that it is indeed becoming increasingly difficult to distinguish between humans and artificial intelligence.

Previously, CAPTCHAs could be easily defeated by algorithms by asking people to identify distorted text. But by 2014, robots had surpassed humans in solving such problems. Many websites have since switched to using image recognition, such as Google, which often asks users to find intersections or traffic lights in images. However, artificial intelligence is now also better than humans in solving such problems.

“The software we have today is making it increasingly difficult for humans to do this,” said Jason Polakis, a professor of computer science at the University of Illinois at Chicago. “We need some alternatives, but there are no real plans for them right now.”

All in all, it’s becoming increasingly difficult to distinguish between humans and computers. This could lead to more bots spreading false information online, for example.