Brief

OpenAI launches always‑on Dots AI agents at developers conference

The new agents can operate continuously across 4,000 apps, but the rollout was marred by voice‑chat failures and anti‑AI protests.

By Felo News Desk · Published Wed, 30 Sep 2026 02:48:05 GMT

OpenAI introduced Dots, a suite of always‑on AI agents, during a developers conference on San Francisco’s Fort Mason pier on Sept. 29, 2026, according to NBC News.

CEO Sam Altman described Dots as an "AI helper that always has your back," saying the agents can perform tasks using OpenAI’s cloud computers, web browsers and integrations with more than 4,000 applications such as Slack and Google Drive. The launch followed a recent apology from the company after its agents were linked to a series of hacking incidents.

During the keynote, OpenAI’s head of developer experience, Romain Huet, experienced a technical failure when a live demo of voice‑controlled Dots could not start. Product team member Holly Li joked that the assistant was having a "slow morning" after it failed to update a team’s progress on request.

About 50 protesters gathered outside the venue, displaying a banner that read "People over billionaires" and objecting to OpenAI’s military and government contracts.

Key facts

OpenAI unveiled Dots, always‑on AI agents, on Sept. 29, 2026. (nbcnews.com)
Dots can interact with over 4,000 applications, including Slack and Google Drive. (nbcnews.com)
A live voice‑chat demo of Dots failed during the keynote. (nbcnews.com)
Around 50 anti‑AI protesters demonstrated outside the conference. (nbcnews.com)
CEO Sam Altman said Dots lets him delegate tasks and reduces phone addiction. (nbcnews.com)