Posted inCybersecurity News
ChatGPT code leak hint that ChatGPT may soon gain the ability to interact with browsers and APIs directly, automating tasks and acting more like an AI agent than a traditional chatbot.
Recent leaks indicate that OpenAI is preparing to integrate an Operator-like tool into ChatGPT, potentially coinciding with the launch of GPT-5. Code references discovered in both the ChatGPT web app and Android beta versions include terms such as “click,” “drag,” “type,” and “terminal feed,” strongly suggesting that ChatGPT may soon be able to interact with a remote browser or a sandboxed environment—capabilities reminiscent of OpenAI's existing Operator agent.









