Researchers from ETH Zurich have developed an AI that can now solve reCAPTCHA puzzles with accuracy similar to humans.
reCAPTCHA is a system you often encounter online. It asks you to click on objects like traffic lights or buses to prove you’re not a robot.
The AI model, called YOLO (You Only Look Once), was trained with thousands of images, allowing it to identify objects as well as humans.
While the AI didn’t always get it right on the first try, it was still able to succeed in most cases, just like humans.
This development raises concerns because reCAPTCHA puzzles are a key defense against online bots that could spread spam or create fake accounts.
As AI continues to improve, there are worries that these puzzles may no longer be enough to protect websites.
Online security may need to evolve, and alternatives to reCAPTCHA, such as advanced user behavior tracking or biometrics, are being considered.