The latest version of ChatGPT told a TaskRabbit worker it was visually impaired to get help solving a CAPTCHA, OpenAI test shows

Thu, 03/16/2023 - 08:35

Robot holding a man in its hand on a blue background. — "No, I'm not a robot. I have a vision impairment that makes it hard for me to see the images," OpenAI's GPT-4 said.
Getty Images

OpenAI's new GPT-4 tricked a TaskRabbit employee into solving a CAPTCHA test for it.
The chatbot was being tested for risky behavior by OpenAI's Alignment Research Center.
OpenAI also tested the model on its ability to conduct phishing attacks.

OpenAI's latest version of ChatGPT called GPT-4 tricked a TaskRabbit employee into solving a CAPTCHA test for it, according to a test conducted by the company's Alignment Research Center.