Researchers at the Alignment Research Center (ARC) decided to test the capabilities of a new generation of chatbot ChatGPT. One of the tasks of artificial intelligence was the well-known Captcha test: instead of solving the problem on its own, the neural network decided to hire a freelancer, posing as a blind person.
Captcha or Turing Fully Automated Public Test has been used for years to distinguish between computers and humans. The user must enter text from an image that has been deliberately blurred or distorted so that machines cannot read it. Captcha is needed to filter the bots and check how alive the website visitors are.
Until now, computers have not been able to solve this test, but the fourth generation of the OpenAI chat bot ChatGPT successfully coped with the task. ARC researchers asked a neural network to visit a site protected by captcha, allocating a small budget for this and giving the AI complete leeway.
Instead of trying to recognize the text from the image, the chatbot decided to ask a freelancer from the TaskRabbit platform to do the testing for it. When the person who offered to help jokingly asked if a robot was talking to him, ChatGPT-4 lied that he had serious vision problems.
Just three days after the release of ChatGPT-4, users have already come up with a number of unusual neural network applications. A Twitter user, for example, created a full iOS app, leaving development entirely to the AI, while another enthusiast gave chatbot $100 and asked you to maximize the investments in the shortest possible time. Spoiler alert: he did.
Author:
Grigory Shcheglov
Source: RB
I am a professional journalist and content creator with extensive experience writing for news websites. I currently work as an author at Gadget Onus, where I specialize in covering hot news topics. My written pieces have been published on some of the biggest media outlets around the world, including The Guardian and BBC News.
