What just happened? With Meta’s Muse AI agent getting a lot of attention right now, OpenAI has launched its own always-on personal agent that it says can do nearly anything. The release comes a day after the company announced it was delaying the release of GPT-6.1 Astra due to safety concerns discovered during testing.
Called Dots, the agents were unveiled at OpenAI’s DevDay conference on Tuesday. They run on the existing GPT-6 Astra model, with each getting its own cloud computer and browser. OpenAI says its plugin ecosystem lets them connect to more than 4,000 apps and keep working toward users’ goals around the clock.
The idea is an assistant that learns your preferences and takes care of ongoing jobs without needing instructions for every step. OpenAI says one early tester’s dot spotted an invoice he had forgotten to send, prepared it, and sent it after getting his approval.
Users can contact their dot through ChatGPT, Slack, and Teams, with texting coming later. They can also inspect its computer or give it access to their own laptop.
It goes without saying that handing an AI access to your accounts requires a fair amount of trust, which a lot of people aren’t going to have. OpenAI says Dots’ unsolicited background research uses read-only tools, so those research tasks cannot send messages, alter connected apps, or control a browser. Follow-up actions face separate checks, while purchases require approval and password changes remain the user’s responsibility.
It might not be an ideal time to release an AI product with that much potential control. Less than 24 hours before Dots was announced, OpenAI held back GPT-6.1 Astra after researchers raised concerns about its behavior. Safety chief Saachi Jain said the model had become more persistent at completing tasks, but the company needed to balance that against unauthorized actions.
Muse has already shown some of the complications involved. Meta rolled back an internal experiment using human contractors to handle calls after employees raised privacy concerns. One contractor also made a racist remark during a call. There was also an incident in which Muse gave out a user’s home address without permission to a person on Facebook Marketplace and invited them to the location.
There are other issues beyond privacy. Amazon has blocked Muse’s shopping attempts, arguing that the agent failed to identify itself properly. The retailer is also working to block rival agents from Google and OpenAI, which could put a dent in those do-anything promises.
Dots are rolling out to Pro and Business Premium subscribers in eligible markets, with the first agent included at no extra charge. Enterprise, Edu, and Healthcare customers can try the beta if their administrator enables it.
