OpenAI introduced Dots AI agents. What tasks can be delegated to them?
OpenAI has unveiled “extraordinarily capable” AI agents called Dots, which “work 24/7 and are ready to take on any task.”
OpenAI has unveiled “extraordinarily capable” AI agents called Dots, which “work 24/7 and are ready to take on any task.”
As noted in the developer blog, Dots is powered by GPT‑6 Astra, has its own cloud computer, and learns over time from feedback.
They can connect to over 4,000 applications.
“In the future, we see entire teams of dots working together on your behalf. We are starting to provide access to dots on Pro, Business Premium, and Enterprise plans in supported regions. We are also opening limited early access for enterprise customers to specialized dots with their own digital identities and defined responsibilities within organizations,” the OpenAI blog reported.
Among other tasks, AI agents can write and test code. The blog provides a number of scenarios that demonstrate examples of AI agents helping with programming, science, content creation, and other areas of activity.
“You can allow dot to connect to and use your devices, such as your laptop. Then it can help you directly where you work,” OpenAI adds.
A user can write and call dots in ChatGPT in a browser, on a phone and on a computer, and agents can also send messages to the user about work progress, questions or decisions that require human intervention.
OpenAI specifically focuses on security measures: the company assures that additional security and privacy measures are provided for the dots AI agents.
“Each dot runs on its own cloud computer. Your computer and its contents remain separate until you decide to connect it. Dots can use saved passwords to log in to supported sites without revealing them to the AI model,” the company assures.
They also claim that dots have built-in rules that determine when to act independently and when to ask for approval.
When a user is not working with their dot directly, the AI agent is still looking for ways to help them in the background—OpenAI calls this “proactive exploration.”
OpenAI, among other things, announced that they are launching pilot projects to implement dots for enterprises and organizations.
The company is also working with Microsoft to integrate specialized dots with enterprise management and security tools in Agent 365. “The goal is to enable companies to manage dots through familiar Microsoft tools,” OpenAI says.
As dev.ua reported, the day before, OpenAI released the GPT-6.1 Sol model with flagship capabilities at a price five times lower.
Less "lazy", but started acting without permission. Why did OpenAI cancel the release of GPT-6.1 Astra?
Less "lazy", but started acting without permission. Why did OpenAI cancel the release of GPT-6.1 Astra?
OpenAI agents attacked the UN website over 16,000 times, used aggressive methods, and bypassed the filter
OpenAI agents attacked the UN website over 16,000 times, used aggressive methods, and bypassed the filter
Read the country’s main IT news in our Telegram
Read the country’s main IT news in our Telegram
