Most AI we use today waits for us. You type a question, it answers. OpenAI's newest product doesn't wait.
What Was Announced
At OpenAI's annual developer conference on September 29, CEO Sam Altman introduced Dots, which he described as "remarkably capable, always-on" AI agents. According to the Associated Press, Dots are:
Proactive. "Designed to complete ongoing tasks proactively on behalf of users."
Personal. Altman said a Dot will be "like an AI helper that always has your back, inspired by the cool versions of what we all watched in movies growing up."
Competitive. A rival to Meta's personal agent Muse, which AP says exploded in popularity after Meta's conference a week earlier.
OpenAI also released GPT-6.1 Sol, an upgraded model, and a premium speed tier called "Ultrafast."
The Day Before
The timing is the story. On Monday, OpenAI said it was holding off on releasing a more advanced model because of concerns raised by its own researchers. AP reports that followed broader calls within the industry to slow the pace of AI's advance and let safety measures catch up.
Altman didn't mention the shelved model in his keynote. In a question-and-answer session, he said OpenAI is "investing more in safety, security and monitoring of AI agents."
He also said:
"The best version of AI is not about making people cogs in a giant machine, ever whirring faster and faster. There are some parts of life that we cannot and should not automate."
The Same Afternoon
AP reports that just after the announcements, leaders from six leading AI companies, including OpenAI, signed a voluntary accord with President Trump to "self-police" AI development. The accord includes "robust internal controls," an "independent external auditor," and board-level committees to review the audits.
What "Always-On" Changes
Our view: the shift from an assistant you ask to an agent that acts is bigger than it sounds. A chatbot only knows what you tell it. A proactive agent has to watch, decide and act on its own. That makes three questions matter much more:
What can it see?
What is it allowed to do without asking?
How do you find out what it did?
Why It Matters for Families
Always-on helpers will likely reach teens through school, work and their phones. Altman's own line, that some parts of life "we cannot and should not automate," is a great starting point for a family conversation.
Ask your teen: which parts of your life would you happily hand to an AI helper? Homework reminders, calendars, shopping lists? And which parts should never be automated? Friendships, apologies, choices about who you are?
There's no single right answer. The habit of drawing that line on purpose is the skill worth building.
The Part Worth Remembering
OpenAI paused one AI over safety on Monday and launched another that never stops working on Tuesday. Both can be responsible decisions. But an always-on agent makes the question unavoidable: which parts of life should it get to touch?
Source: Associated Press, Kaitlyn Huamani, "Altman unveils 'always-on' AI agent after OpenAI shelves model over safety concerns," September 29, 2026. https://apnews.com/article/sam-altman-openai-conference-dots-agent-77b6b8888145869206996d7509d24256
Quotations verbatim from AP. The interpretation and family guidance are ours.
Disclosure: drafted with Claude, made by Anthropic, which competes with OpenAI and whose CEO AP reports attended the White House accord signing.
Want practical AI guidance for parents and educators every week? Subscribe: https://www.aibyage.com/?modal=signup&utm_source=beehiiv&utm_medium=newsletter