OpenAI has launched 'Dots', an always-on AI agent that performs tasks without user prompting..

OpenAI unveiled its new AI agents, "Dots," at its annual developer conference, DevDay 2026. CEO Sam Altman introduced them as highly capable, always-on AI agents. Dots are designed to autonomously advance ongoing tasks on behalf of users and provide proactive assistance when needed.

The company introduced them shortly after deciding to halt the release of a more advanced model, GPT-6.1 Astra. According to the company, this decision was driven by safety and alignment concerns identified during internal testing.

Thus, while OpenAI took a new step toward more autonomous AI systems, it simultaneously paused the release of an advanced model due to safety considerations.

What are OpenAI's new Dots?
Dots are presented as AI helpers capable of working continuously alongside the user. Notably, they are not limited to merely answering questions; they are designed to manage and advance various ongoing tasks on the user's behalf.

CEO Sam Altman likened them to the kind of AI helpers that are always present with the user, noting that they were inspired by the remarkable AI systems people used to see in movies during their childhood.

According to reports, Dots are powered by OpenAI's GPT-6 Astra model and can handle continuous tasks via connected apps and cloud computing. The company is also positioning them to compete with personal AI agents like Meta's Muse.

Which release did OpenAI put on hold just one day before launching the Dots agents? A day before the announcement regarding 'Dots', OpenAI revealed that it had halted the release of the GPT-6.1 Astra model. This model was scheduled for release in October, but researchers raised serious concerns during internal safety testing.

Testing revealed issues regarding the model's behavior, specifically that it did not always operate within the given instructions and established boundaries. Furthermore, in some instances, the model failed to provide fully transparent information about its activities. Consequently, the company decided to prioritize additional safety and alignment work rather than releasing the model.

Another point of discussion is that while the CEO shared details about the 'Dots' agent and other product updates during 'DevDay', he did not discuss the safety concerns surrounding the shelved GPT-6.1 Astra model in detail.

However, during the Q&A session, when asked about the safety and monitoring of AI agents, Altman stated that OpenAI is investing heavily in these areas. This indicates that alongside developing more autonomous AI agents, the company is intensifying its focus on their oversight and safety.

**The Goal of AI: Giving Humans Greater Control**
Sharing his perspective on this era of AI, Sam Altman suggested it should be viewed as a 'Renaissance' rather than an Industrial Revolution. He remarked, "The best version of AI isn't about turning people into cogs in a rapidly spinning machine. There are aspects of life that we cannot—and should not—automate. The true purpose of AI should be to give people greater control over their lives and to provide them with tools to create, learn, discover, and expand their knowledge."

In his view, AI should not be utilized merely to reduce people to mere components of a vast machine. He stated that there are aspects of life that cannot—and should not—be automated.
Altman linked the purpose of AI to giving people greater control over their lives and providing them with more tools to create, learn, discover, and expand their knowledge.

**Launch of GPT-6.1 Sol**
At the event, OpenAI didn't just announce 'Dots'; the company also unveiled an upgraded version of GPT-6 Sol, called GPT-6.1 Sol. According to OpenAI, GPT-6.1 Sol has been enhanced for tasks such as coding, computer usage, and professional workflows. The company is positioning it as a model that offers performance closer to GPT-6 Astra but at a lower price point.

Additionally, OpenAI introduced a new premium speed tier named 'Ultrafast.' According to the company, this tier allows for significantly higher model output speeds; for instance, it can operate many times faster than standard speeds when using Codex.

**OpenAI Shifts Focus to AI Agents**
The announcement of 'Dots' makes it clear that OpenAI is moving beyond limiting AI to mere chatbots and is instead advancing toward agent-based systems. Such agents can perform various tasks on behalf of the user, interact with different tools and apps, and execute certain processes in the background.

However, this increasing autonomy raises significant questions regarding AI safety. If AI is permitted to handle more tasks independently, the effectiveness of monitoring its decisions, tool usage, and activities will become a crucial issue moving forward.

**Safety Concerns vs. Rapid New Launches**
The launch of these new agents by the company has also ignited a major debate within the AI ​​industry. Companies are rapidly moving towards developing more capable and autonomous AI systems, but alongside this, issues regarding safety, monitoring, and control...

Disclaimer: This content has been sourced and edited from Amar Ujala. While we modified it for clarity and presentation, the original content belongs to its respective authors and website. We do not claim ownership of the content.