ChatGPT Agent was able to pass the "I am not a robot" captcha prevention on a website. The internet can't believe what happened.
For those unaware, OpenAI launched its ChatGPT Agent a few weeks ago. It can automate tasks that you set, from start to finish. It's only available for ChatGPT Pro users for now.
A reddit user shared some screenshots that showed how ChatGPT Agent was narrating through the task it was doing. When it came across a Cloudflare captcha, ChatGPT Agent said "The link is inserted, so now I'll click the 'Verify you are human' checkbox to complete the verification on Cloudflare. This step is necessary to prove I'm not a bot and proceed with the action."
Some people were shocked that an AI could bypass captchas so easily, especially when humans have trouble with a few. Others joked about the irony of a chatbot passing the captcha successfully to prove it was human. Technically, ChatGPT is an LLM, as users have pointed out. It's not a bot, it has been trained on human usage, such as clicking on things with the cursor. Speaking of which, it is worth noting that this was a regular click captcha, and not one that requires you to click on a few images.
That's what captchas are designed for, right? To stop automated clicks? Not really, on the face captchas are challenges that require users to prove they are not a robot. But, Ars Technica notes they are screening systems that analyze various things such as mouse movements, the time a user takes to click, browser fingerprint, JavaScript execution patterns, etc. Often, the reason you may run into captcha checks is because of IP reputation. When the captcha detects something unusual, it blocks the user from proceeding by placing a visual challenge.
You might think that if ChatGPT Agent can bypass captchas (which it didn't, it PASSED the test), what's to stop hackers from using it for attacking websites? This is why captchas are used, not to block users, but to protect websites from bot attacks like a distributed denial-of-service attack (DDoS). Actually, those are just scripts that spam HTTP requests. An attacker wouldn't need an AI for such attacks, besides LLMs require a lot of resources, and are not cheap. Using AI to craft malware, well, that's a different story.
If anything, this incident shows that captchas need to evolve. Some users say that ChatGPT agent was not able to solve captchas, it's not supposed to in the first place.
Thank you for being a Ghacks reader. The post OpenAI's ChatGPT Agent passes "I am not a robot" captcha appeared first on gHacks Technology News.
0 Commentaires