OpenAI is reportedly building a new AI agent tool that can directly carry out tasks on a user’s computer. The purported new tool, internally referred to as "Operator," could be launched in January 2025, according to a Bloomberg report. It has been designed as a research preview tool and will be available to developers through OpenAI’s application programming interface (API). As per the report, Operator will be rolled out as a general-purpose AI tool with the capability of navigating the Internet through a web browser. OpenAI’s reported offering comes after Anthropic recently debuted a similar AI agent feature called "computer use" as part of its latest iteration of Claude AI models. With Claude 3.5 Sonnet, developers can hand over control to virtual AI assistants capable of interacting with any desktop app through mouse clicks, keystrokes, etc. Google is also rumoured to be building an AI agent product with a focus on consumers. It is part of an industry-wide shift to automate back-office tasks such as answering emails, filling out forms, etc. using AI-powered virtual assistants. A recent survey by Capgemini revealed that over 82 per cent of organisations are planning to integrate AI agents within the next three years for automating email generation, coding, data analysis, and more.
