Friday, January 24, 2025

OpenAI Gives Sneak Peek Into Its First AI Agent ‘Operator’ With Launch Due Soon

OpenAI promised that 2025 would be a major year for AI agents and it seems like the company is staying true to that.

In case you’re still wondering, these are tools that can conduct tasks automatically. We mean they can even perform human-like actions on their own when instructed to do so. OpenAI shared a sneak peek at what many users can expect from its Operator agent as it hopes for a launch in the US soon.

The general purpose agent takes control of web browsers and conducts certain tastes independently. It would be up for grabs to those paying $200 a month for Pro subscriptions first which does make sense. However, the agent will soon be a part of the Plus, Enterprise, and Team tiers.

Altman shared how he plans on a global launch as well but a launch in the EU might take some time, thanks to the greater laws related to the tech world involved. Those wanting a little extra information can head on over to operator.chatgpt.com for the research preview. With time, the AI agent would soon become part of its host of clients on ChatGPT.

So far, tasks that can be performed with ease by the Operator include making travel bookings, reservations at restaurants, and carrying out online shopping tasks. There’s a list of categories that users can select from such as dining, shopping, travel, and delivery. All of these enable various types of automation.

When the Operator is activated, a tiny window pops up that displays dedicated web browsers that agents use for fulfilling tasks, alongside explanations for certain actions that they would be performing. Users could still take full command of their screen if and when the Operator is in action. This is related to the fact that the Operator has a separate dedicated browser.

As per OpenAI, the tool is powered using CUA that merges visual services with the firm’s GPT-4o model. This also entails reasoning abilities in more advanced systems. The agent can interact more with websites which means it does not need to make use of APIs facing developers to perform tasks.

The CUA makes use of various navigation menus, buttons, and options for filling forms on pages like we do in the regular world. Soon, the company also hopes to collaborate with other leading firms such as Instacart, StubHub, Uber, and Priceline to make way for seamless partnerships that respect their service agreements.


Read next: AI-Driven Changes Will Reshape Mobile Engagement, Social Media Budgets, and Consumer Data Collection by 2027
by Dr. Hura Anwar via Digital Information World

No comments:

Post a Comment