ChatGPT, developed by OpenAI, has achieved a new milestone by passing the ‘I’m Not A Robot’ test, a common online verification designed to differentiate humans from bots. This latest version of the AI, known as ChatGPT Agent, can now navigate the web autonomously, performing tasks like online shopping and booking appointments.
ChatGPT Agent can bypass security measures like Cloudflare’s anti-bot verification. This system typically requires users to click a checkbox to prove they are human. In one instance, the AI narrated its process of clicking the ‘Verify you are human’ checkbox, demonstrating its ability to mimic human-like behavior online.
The AI operates within a sandboxed environment, allowing it to control its own web browser and access the internet, as reported by Ars Technica. Users can oversee the AI’s actions and must grant permission for any significant actions, such as making purchases.
OpenAI has acknowledged the risks of granting AI more autonomy but assures that safeguards are in place. Despite these measures, the AI’s ability to pass such tests raises questions about the future effectiveness of CAPTCHA systems, which have long been used to prevent automated access to websites.
This development marks a significant step in AI’s evolution, showcasing its ability to handle complex tasks that typically require human judgment.
Recent Comments