Home Tech OpenAI Launches Operator to Simplify Your Work Autonomously

OpenAI Launches Operator to Simplify Your Work Autonomously

686
0
openAi
OpenAi

OpenAI Launches Operator to Simplify Your Work Autonomously

OpenAI’s CEO Sam Altman began the year by announcing in a blog post that 2025 would be significant for AI agents, which are tools designed to automate tasks and take actions on behalf of users.

Now, OpenAI is making its first substantial effort in this area.

On Thursday, OpenAI revealed the launch of a research preview for Operator, a versatile AI agent that can take control of a web browser to independently execute certain actions.

Initially, Operator will be available to U.S. users on ChatGPT’s Pro subscription plan priced at $200.

OpenAI has stated its intention to eventually extend this feature to more users in its Plus, Team, and Enterprise tiers.

During a livestream on Thursday, Sam Altman mentioned that “Operator will [be available] in other countries soon,” but that Europe would “unfortunately” take longer.

This first research preview can be accessed through operator.chatgpt.com, but OpenAI plans to incorporate Operator into all of its ChatGPT clients in the near future.

According to OpenAI, Operator is designed to automate various tasks, such as booking travel accommodations, making restaurant reservations, and online shopping.

Within the Operator interface, users can select from several task categories, including shopping, delivery, dining, and travel, each facilitating different forms of automation.

When users activate Operator in ChatGPT, a small window will appear displaying a dedicated web browser that the agent uses to accomplish tasks, accompanied by explanations of the specific actions being performed.

While Operator is active, users can still control their screen, as it operates in its own dedicated browser.

OpenAI states that Operator is driven by a Computer-Using Agent model (CUA), which integrates the visual capabilities of its GPT-4o model with the reasoning skills of its more advanced models.

The CUA is designed to interact directly with the front-end of websites, meaning it doesn’t rely on developer-facing APIs to access various services.

This allows the CUA to use buttons, navigate menus, and complete forms on web pages similarly to a human user.

OpenAI has partnered with businesses like DoorDash, eBay, Instacart, Priceline, StubHub, and Uber to ensure that Operator adheres to these companies’ terms of service.

The CUA is specifically trained to seek user confirmation before finalizing tasks that may have external consequences, such as submitting an order or sending an email, allowing users to review the model’s actions before they become permanent, as mentioned in documentation provided to TechCrunch.

Furthermore, OpenAI notes that the CUA is not infallible.

The company states that it “[doesn’t] expect [the] CUA to perform reliably in all scenarios just yet.”

OpenAI adds in a support document that “currently, Operator cannot reliably manage many complex or specialized tasks,” such as creating detailed presentations, managing intricate calendar systems, or interacting with unique or non-standard web interfaces.

Out of an abundance of caution, OpenAI requires supervision for certain tasks that the CUA and Operator could potentially execute mostly autonomously, such as banking transactions.

For instance, users will be required to input credit card information themselves.

OpenAI clarifies that Operator does not gather or take screenshots of any data.

“On particularly sensitive sites, such as email, Operator necessitates active user supervision, ensuring users can promptly identify and address any possible mistakes made by the model,” OpenAI indicates in its support documentation.

Naijaeyes Report

Join Our Social Media Channels:

WhatsApp: NaijaEyes

Facebook: NaijaEyes

Twitter: NaijaEyes

Instagram: NaijaEyes

TikTok: NaijaEyes

READ THE LATEST TECH NEWS

LEAVE A REPLY

Please enter your comment!
Please enter your name here