OpenAI's New 'Dots' Promise Proactive AI — But Most ChatGPT Users Are Left Waiting

There's a buzz in the air, isn't there? OpenAI, the folks behind ChatGPT, have just pulled back the curtain on something truly exciting – a new feature they're calling 'Dots.' Now, before you rush off to check your ChatGPT account, there's a pretty big catch: chances are, you won't find them there. Not yet, anyway. It seems this groundbreaking step towards truly proactive AI is, for now, an exclusive club.

So, what exactly are these 'Dots' we're talking about? Imagine ChatGPT, but with a serious upgrade in initiative. We're not just talking about answering your prompts anymore; 'Dots' are designed to be proactive, almost like a tireless digital assistant that keeps working towards your goals even when you've logged off for the day or stepped away from your keyboard. Think of them as persistent, goal-oriented AI agents, always chipping away at the tasks you've set for them.

The magic behind 'Dots' is quite fascinating. Each Dot essentially gets its own cloud-based computer and a dedicated browser. This means they can perform tasks, conduct research, and interact with the digital world independently. And here's where it gets truly powerful: they integrate with a massive ecosystem – over 4,000 apps, thanks to OpenAI's robust plugin system. You can set them on a task, and they'll get to work. Want to check in or give new instructions? You can do that right through ChatGPT, or even your existing Slack and Teams channels. Pretty neat, right?

Underpinning this impressive capability is OpenAI's latest linguistic marvel, the GPT-6 Astra model. Now, for the less exciting news – the availability. As of their September 29, 2026 launch, 'Dots' are a bit of an exclusive club. They've been rolled out initially to ChatGPT Pro and Business Premium subscribers in what OpenAI deems 'eligible markets.' If you're an enterprise user – think big companies, educational institutions, or healthcare providers – you might be able to get your hands on a beta version, provided your workspace administrator has enabled it. But for the vast majority of us, those on Plus, Go, or the free ChatGPT plans? We're currently out of luck. OpenAI promises wider access 'soon,' but as we all know, 'soon' in tech often means 'when it's ready and maybe a bit pricey.'

You might be wondering, why the slow rollout? Well, features this powerful, this autonomous, naturally come with a lot of complexity – and responsibility. OpenAI has put a significant emphasis on safety, and frankly, that's reassuring. When a Dot is doing its 'proactive research' in the background, its access to connected-app tools is strictly read-only. This means it can't go rogue, messaging people, changing your content, or taking control of your browser or computer without explicit permission. Imagine the potential for mischief if that wasn't the case!

Users remain firmly in control, which is paramount. You decide which apps your Dot can access, and you can even set 'Custom Rules' for actions that either require your direct approval or are blocked altogether. There’s also an 'auto-review' system diligently checking every action that might affect your accounts or share information, making sure it aligns with your instructions and OpenAI’s stringent safety requirements. And for truly sensitive actions, like, say, changing a password? That will always demand your direct input. No AI agent, however smart, is getting near your banking details without you. This meticulous approach has seemingly paid off during testing of the GPT-6 Astra model, showing a mere 3% rate of 'misaligned outcomes' – a significant improvement from the 8% seen with its predecessor, GPT-5.6 Sol.

Now, for the inevitable question: what about the cost? Running sophisticated, proactive AI agents in the cloud sounds like it could be computationally intensive, and thus, not cheap. OpenAI seems to be addressing this by including the first Dot as part of the Pro and Business Premium plans. Conversations with your Dot won't count against your regular ChatGPT usage limits, and you get an allowance for 'deeper work.' This certainly suggests that for those who want to add more Dots or expand their 'work capacity,' there will likely be additional costs down the line. It's a premium feature, after all, and powerful tools often come with a price tag.

While many of us are still patiently waiting for our turn, 'Dots' represent a fascinating leap forward for AI, moving it closer to true agentic capability. It's exciting to imagine an AI that truly works for you, even when you're not actively prompting it. For now, the select few can enjoy this glimpse into the future, and the rest of us will just have to keep an eye out for that much-anticipated 'soon.'