OpenAI Operator: A Step Towards Autonomous AI | OpenAI italiano | Chat GPT gratis | OpenAI Playground | Turtles AI

OpenAI is preparing to launch Operator, an advanced AI agent designed to autonomously handle complex tasks. Leaked benchmarks reveal its strengths and weaknesses, as competition heats up with other tech giants like Anthropic and Google.

Key Points:

• Operator: OpenAI’s new AI agent, designed to handle complex tasks.
• Security and challenges: Tests show promising performance but not yet complete reliability.
• Competition: Anthropic and Google prepare alternatives to challenge OpenAI.
• Upcoming innovation: Autonomous AI approaches a central role in technology.

OpenAI is about to introduce a new frontier in AI with the launch of Operator, an agent designed to automate complex tasks beyond the capabilities of existing technologies. The tool is rumored to be capable of everything from writing code to booking travel, promising to autonomously manage users’ daily activities. Engineer Tibor Blaho, known for his precise anticipations in the industry, has identified tangible evidence of Operator’s existence within the code of ChatGPT’s macOS client. Among the hidden features are options like “Toggle Operator” and “Force Quit Operator,” clear signs of an advanced project. In parallel, OpenAI’s website contains non-public references that show Operator’s comparative benchmarks with other systems, highlighting the platform’s progress as well as limitations.

Operator’s performance has been measured through several tests, including OSWorld, which simulates a real-world computing environment. In this context, the system achieved a score of 38.1%, higher than its competitors but still far from the human level of 72.4%. Better performance was recorded in WebVoyager, a platform that evaluates interaction with websites, while in WebArena, another web browsing test, the model did not achieve results comparable to those of humans. The most obvious challenges emerge in practical tasks: Operator’s success in registering with a cloud provider and launching a virtual machine stands at 60%, while it managed to create a Bitcoin wallet in only 10% of cases.

Despite the shortcomings, OpenAI has placed a strong emphasis on the security of the system. According to leaked documents, Operator has proven to effectively resist attempts to exploit it for illicit activities or search for sensitive data. The long development of this agent is attributable precisely to the focus on security, an aspect that OpenAI co-founder Wojciech Zaremba has emphasized, criticizing competitors such as Anthropic for releasing less protected systems.

The competition is growing: Anthropic, Google, and other industry players are accelerating the launch of their autonomous agents. Google is expected to unveil its system by December, while OpenAI is targeting January for a preliminary version of Operator, initially aimed at developers and researchers. These AI agents, while still primitive, represent a huge step forward in technological autonomy, so much so that OpenAI CEO Sam Altman has called agents the next big evolution of AI.

The market for this new technology is estimated to be $47.1 billion by 2030, according to Markets and Markets, and it promises to redefine the relationship between humans and machines. OpenAI’s progress, combined with growing competition, means that 2025 could be a big year for the introduction of AI that can improve the way we interact with technology.

With Operator, OpenAI is redefining the AI ​​landscape, opening up new possibilities and challenges for the future.