OpenAI has launched its latest AI agent, “Operator,” designed to perform digital tasks on behalf of users. However, the experimental technology is still in its infancy and requires “a lot of adult supervision.” Bloomberg reporter Rachel Metz shared her experience with Operator, describing it as slow and sluggish, frequently asking for help when stuck.
The AI model’s capabilities are limited by its current development stage. While it can browse the web and interact with desktop environments, it lacks the reliability needed to serve as a virtual employee or personal assistant. OpenAI highlights Operator’s usefulness in tasks like making reservations and creating shopping lists.
However, concerns remain about Operator’s trustworthiness. The AI model must ask for user confirmation before completing important tasks, indicating it is not yet autonomous enough. Bloomberg’s Metz found success with smaller tasks like ordering ice cream, but struggled with more complex applications such as creating spreadsheets.
The high price point of $200 per month for ChatGPT Pro plan users may make Operator less appealing to some. To take AI agents to the next level, OpenAI must convince users that they can trust these services to act autonomously on their behalf. Until then, relying on human intervention remains the best option.
Source: https://futurism.com/the-byte/openai-ai-agent-monitored-nonstop