CAPTCHAs are suddenly so difficult—are you struggling?
- Replies 1
If you've recently found yourself squinting at your screen, scratching your head, and questioning your humanity while trying to decipher a Captcha, you're not alone.
The once straightforward task of proving you're not a robot has become frustrating for many, and these puzzles are only getting harder.
But why is this happening?
CAPTCHAs, which stand for 'Completely Automated Public Turing test to Tell Computers and Humans Apart,' have been a staple of internet security since their inception in 2000.
Their primary purpose is to prevent automated bots from spamming websites, scraping content, or skewing online polls.
Initially, CAPTCHAs were simple—a jumble of distorted letters and numbers you had to type out.
But as technology has advanced, so too have the bots, leading to an arms race between CAPTCHA designers and bot developers.
The evolution of CAPTCHAs has been significant.
From the early days of distorted text, we've moved on to image recognition tasks, such as identifying traffic lights, crosswalks, or buses in a grid of photos. These tasks were designed to be easy for humans but challenging for bots.
However, as artificial intelligence and machine learning have become more sophisticated, bots have learned to solve these puzzles with alarming accuracy.
In response, CAPTCHA designers have had to up the ante.
The latest iterations involve more complex tasks, such as aligning objects of the same shape or selecting non-aquatic animals from a lineup.
‘Things are going to get even stranger, to be honest because now you have to do something that's nonsensical,' Kevin Gosschalk, CEO of web security firm Arkose Labs, pointed out.
'Otherwise, large multimodal models will be able to understand.’
The goal is to create tasks beyond the current capabilities of bots, but unfortunately, this also means they're becoming more difficult for humans.
'I've always hated the “click the traffic lights” or “crossing ones” because I'm never sure if it means all of it,' one person complained on social media.
'If it's 95 per cent in one square, is that both squares?'
In identifying whether a bot is a low, medium, or high-risk to a site, Arkose MatchKey Labs tailored the difficulty of the CAPTCHA to the perceived threat level of bots.
They start with an ‘easier’ option where you have to use arrows to match a crab on the left to a bear on the right.
But things get trickier when you visit a high-risk website. Now, you have to ‘use the arrows to change the number of objects until it matches the left image’.
The company claims to have the ‘toughest CAPTCHA ever made’.
Most websites only approve it if you can solve the puzzle on your first try. However, they admit that for the toughest puzzles ‘designed for bad actors’, the success rate of users doesn't affect how well it works.
As bots have gotten better at cracking CAPTCHAs, the codes have had to become more difficult.
Now, instead of just distorted text, users might be asked to solve a simple math problem.
Last year, researchers at the University of California, Irvine discovered that bots could solve CAPTCHAs with distorted text with almost 100 per cent accuracy.
Bots are often made to scrape content from websites, post fake comments or reviews, and they ‘often outsource solving to CAPTCHA farms—sweatshop-like operations where humans are paid to solve CAPTCHAs,’ the study said.
The increasing complexity of CAPTCHA has even caught the attention of celebrities like British comedian Jack Whitehall, who joked about the absurdity of failing these tests in his Netflix special.
'Is it just me, or have those “I am not a robot” tests started getting harder?' he joked.
'Has anyone had that moment recently where you have failed the I-am-not-a-robot test so many times that you have that moment where you stop and go…Maybe I am a robot?’
'I haven't been able to spot 10 [stop]lights in a row. I'm either a robot or a cyclist!'
Have you found CAPTCHAs a nuisance, or do you see them as necessary? Share your thoughts in the comments below!
The once straightforward task of proving you're not a robot has become frustrating for many, and these puzzles are only getting harder.
But why is this happening?
CAPTCHAs, which stand for 'Completely Automated Public Turing test to Tell Computers and Humans Apart,' have been a staple of internet security since their inception in 2000.
Their primary purpose is to prevent automated bots from spamming websites, scraping content, or skewing online polls.
Have you also struggled with CAPTCHAs? Credit: Shutterstock
Initially, CAPTCHAs were simple—a jumble of distorted letters and numbers you had to type out.
But as technology has advanced, so too have the bots, leading to an arms race between CAPTCHA designers and bot developers.
The evolution of CAPTCHAs has been significant.
From the early days of distorted text, we've moved on to image recognition tasks, such as identifying traffic lights, crosswalks, or buses in a grid of photos. These tasks were designed to be easy for humans but challenging for bots.
However, as artificial intelligence and machine learning have become more sophisticated, bots have learned to solve these puzzles with alarming accuracy.
In response, CAPTCHA designers have had to up the ante.
The latest iterations involve more complex tasks, such as aligning objects of the same shape or selecting non-aquatic animals from a lineup.
‘Things are going to get even stranger, to be honest because now you have to do something that's nonsensical,' Kevin Gosschalk, CEO of web security firm Arkose Labs, pointed out.
'Otherwise, large multimodal models will be able to understand.’
The goal is to create tasks beyond the current capabilities of bots, but unfortunately, this also means they're becoming more difficult for humans.
'I've always hated the “click the traffic lights” or “crossing ones” because I'm never sure if it means all of it,' one person complained on social media.
'If it's 95 per cent in one square, is that both squares?'
In identifying whether a bot is a low, medium, or high-risk to a site, Arkose MatchKey Labs tailored the difficulty of the CAPTCHA to the perceived threat level of bots.
They start with an ‘easier’ option where you have to use arrows to match a crab on the left to a bear on the right.
But things get trickier when you visit a high-risk website. Now, you have to ‘use the arrows to change the number of objects until it matches the left image’.
The company claims to have the ‘toughest CAPTCHA ever made’.
Most websites only approve it if you can solve the puzzle on your first try. However, they admit that for the toughest puzzles ‘designed for bad actors’, the success rate of users doesn't affect how well it works.
Arkose MatchKey Labs tweaked its CAPTCHA to perceive the risk level of bots. Credit: Arkose Matchkey Labs
As bots have gotten better at cracking CAPTCHAs, the codes have had to become more difficult.
Now, instead of just distorted text, users might be asked to solve a simple math problem.
Last year, researchers at the University of California, Irvine discovered that bots could solve CAPTCHAs with distorted text with almost 100 per cent accuracy.
Bots are often made to scrape content from websites, post fake comments or reviews, and they ‘often outsource solving to CAPTCHA farms—sweatshop-like operations where humans are paid to solve CAPTCHAs,’ the study said.
The increasing complexity of CAPTCHA has even caught the attention of celebrities like British comedian Jack Whitehall, who joked about the absurdity of failing these tests in his Netflix special.
'Is it just me, or have those “I am not a robot” tests started getting harder?' he joked.
'Has anyone had that moment recently where you have failed the I-am-not-a-robot test so many times that you have that moment where you stop and go…Maybe I am a robot?’
'I haven't been able to spot 10 [stop]lights in a row. I'm either a robot or a cyclist!'
.png){kind=icon}
