BOFFINS HAVE created a neural network smart enough to fool Captcha technology.
Captcha is that irritating thing where you (usually) have to type in two words on the screen to prove you're not a robot. The best-known is Google's reCaptcha which has progressed from words to being able to tell from a simple box tick if you're real or not.
It's not even the first time Captcha has been hacked this year alone, but Science reports (via BBC), a team from Vicarious - co-funded by Jeff Bezos (Amazon) and Mark Zuckerberg (Facebook) - has found a way to beat the text-based version of Captcha using artificial intelligence (AI).
Given that Captcha is an acronym for "Completely Automated Public Turing test to tell Computers and Humans Apart", that's a bit awkward.
Technically, Vicarious is a Recursive Cortical Network (RCN) rather than a traditional neural network. RCN's are much closer to how you and I think and take a lot less processing power.
The RCN has now learned to crack Captcha used on major sites including Google, Yahoo, PayPal, and the Captcha website with up to 90 percent accuracy.
Those sites have since upped their Captcha game, but the tech which learns to determine shapes from their pixel outline can still manage between 57 and 66 per cent accuracy.
As artificial intelligence improves, we're likely to see more and more cases like this. As Simon Edwards from cyber-security experts Trend Micro Europe told the BBC, "The very nature of big data analysis and machine learning is that if you give it enough data to play with, it will eventually work out most things."
He goes on to recommend two-factor authentication as the most foolproof way of protecting accounts. This usually means partnering a password with something else, such as biometrics, a text message with a second code, a generator like those used by some banks, or a FIDO key like these from Yubico, which is what we tend to use. µ
The week in Google
The scandal that just keeps giving
Clip to the end....