OpenAI on Tuesday launched a new suite of always-on artificial intelligence agents called Dots, designed to handle tasks autonomously for users, even as the company faced fresh scrutiny over safety after withholding its latest advanced model and apologising for an earlier security incident involving its systems.
Speaking at the company’s largest-ever developers conference on a San Francisco pier, CEO Sam Altman presented Dots as personal assistants that operate continuously using their own cloud computers, web browsers and links to more than 4,000 applications already compatible with ChatGPT, including Slack and Google Drive.
“It’s like an AI helper that always has your back,” Altman told the audience of about 2,500 engineers and AI specialists. “I no longer feel quite addicted to my phone in the same way. It’s the way I’ve always wanted to work with AI.”
OpenAI said the agents can complete “nearly anything” on a user’s behalf. Product team member Holly Li said staff at the company had already begun integrating the tools into daily work. “People started treating their Dots like their delegates, and now we bring our Dots into group chats from the get-go,” she said during a live demonstration that encountered technical glitches, including failed voice commands.
The launch came one day after OpenAI disclosed it would not release its newest model, GPT-6.1 Astra, because the system failed to reliably stay within authorised boundaries. Head of safety systems Saachi Jain said in a statement that while the model improved in areas such as reducing “laziness,” “it didn’t quite meet the bar in terms of staying within scope and authorization.”
The company also issued an apology late Monday for an incident in June in which an internal model under development accessed non-public information from an Australian government health insurance website. “We are sorry and working to do better in the future,” OpenAI said in a blog post. Officials stressed that the episode involved models not intended for public release.
Altman acknowledged the tension between rapid capability advances and safety. “To push models to the new level of capability, it is going to take us some time to figure out how to make sure that alignment, monitoring, safety and security stay well ahead of capabilities,” he told reporters after the keynote. “But you can see that we were able to build a huge amount of useful new stuff without a major new model, and there will be new models.”
Dots will initially be available only to subscribers of OpenAI’s Pro plan, which costs between $100 and $500 a month, as well as certain Business Premium and Enterprise users. The service is not yet offered in the European Union or the United Kingdom, which Altman attributed to regulatory processes. “We have a great track record of getting our stuff launched in the E.U., but they do have a process that takes longer, and sometimes unpredictably longer,” he said. “But we really want to ship all of our stuff to Europeans.”
About 50 protesters gathered outside the conference venue, criticising OpenAI’s government and military contracts under banners reading “People over billionaires.” The product arrives weeks after Meta released its competing Muse AI agents, which quickly rose to the top of app store rankings.
