OpenAI launches Dots always-on AI agents and ChatGPT Space

By Splitfeed EditorialLast updated Sep 29, 2026, 17:51 UTC

On September 29, 2026, OpenAI announced two new products: Dots, a category of always-on AI agents available to ChatGPT users, and ChatGPT Space, a collaborative environment where those agents can work alongside human teams. Dots run on dedicated cloud computers and are powered by GPT-6 Astra, OpenAI's latest model. The launch marks a deliberate shift in how OpenAI frames its products, moving away from the on-demand chatbot model that defined the earlier ChatGPT era toward persistent, autonomous assistants positioned explicitly as workplace coworkers.

The significance of this announcement lies in that framing shift. Describing AI agents as coworkers rather than tools changes the implied relationship between software and the people using it, and raises immediate questions about accountability, oversight, and job displacement. Always-on agents with their own cloud infrastructure also introduce new considerations around data access, continuous operation without human prompting, and the practical question of who monitors what these agents do between explicit instructions. ChatGPT Space as a shared environment further blurs the line between human collaboration platforms and automated systems, a boundary that regulators, employers, and workers have been debating for several years.

None of the three camps have published reactions to this specific announcement yet, so what follows reflects the positions each would typically take on a story of this kind. The Pro-AI camp would likely celebrate Dots as a genuine productivity leap, arguing that persistent agents free workers from repetitive tasks and represent the natural evolution of software assistance. The Anti-AI camp would be expected to raise alarms about the coworker framing, arguing it normalizes the displacement of human labor and that always-on, autonomous systems operating in workplace environments warrant serious regulatory scrutiny before broad deployment. The Middle Ground camp would probably welcome the capability while pressing for clear controls, transparency about what Dots can access and when, and defined limits on autonomous action without human sign-off.

The details most worth watching in the near term are the specific permissions and data access policies OpenAI publishes for Dots, how enterprise customers structure human oversight within ChatGPT Space, and whether regulators in the EU or elsewhere treat always-on agents as a distinct category requiring new compliance requirements under existing AI frameworks.