OpenAI’s CEO, Sam Altman, just described a version of ChatGPT that watches everything you do, and the internet has plenty of feelings about it. Speaking at Internapalooza in a conversation with Z Fellows founder Cory Levy, Sam Altman laid out a near-future vision for ChatGPT that goes well beyond answering questions in a chat box.

What Sam Altman Actually Said

Altman described a ChatGPT descendant capable of watching a computer screen at all times, sitting through every meeting, recording every call, and carrying full context of a person’s entire life. Crucially, he framed this as opt-in and user-controlled — people would choose what the assistant could access, connecting services like text messages, email, documents, or Slack themselves, with the AI not making decisions on their own behalf.

The most-quoted line from the interview was his timeline estimate: Altman said he thinks OpenAI is only about one model generation away from this actually being genuinely useful.

OpenAI and Hugging Face partner to address security incident during model evaluation

What This Assistant Would Actually Do

Capability	Detail
Screen monitoring	Continuous, always-on observation
Meeting/call tracking	Records and retains meeting/call context
Data connections	User-chosen: texts, email, docs, Slack
Decision-making	Explicitly does not make decisions for the user
Core function	Catches mistakes, suggests ideas, handles overflow tasks
Timeline	~6 months for initial capability, one model generation for real usefulness

Altman pitched the use case around overloaded professionals — the example he gave was a startup CEO drowning in more tasks than any one person can handle, with this kind of assistant working alongside them rather than replacing their judgment.

Not Exactly Unprecedented Territory

This isn’t OpenAI’s first step in this direction. ChatGPT’s existing memory feature already lets it reference past conversations for more tailored responses, with users able to turn saved memories off in settings. Microsoft has also already tested something structurally similar with Windows Recall, a Copilot+ PC feature that snapshots screen activity — though only after facing enough backlash to add opt-in requirements, local encryption, and Windows Hello authentication before its eventual rollout.

The Reaction Has Been Loud

Predictably, the response to Altman’s comments has skewed heavily skeptical, with plenty of public figures and commentators openly mocking the idea of an AI system with that level of access to their daily lives. The core tension is straightforward: an assistant that genuinely needs full context to be maximally useful is, by definition, an assistant that needs comprehensive access to a person’s private communications and activity — a trade-off plenty of people aren’t eager to make regardless of how “opt-in” the framing is.

You can read more about OpenAI’s product direction on the official OpenAI blog.

For more on how AI companies are pushing the boundaries of personal data access, check out TechnoSports’ coverage of Google’s optional AI watermark removal and the broader AI industry landscape shaping 2026.

Bottom Line

Whether this ships in six months or takes considerably longer, Altman’s comments make OpenAI’s ambitions for ChatGPT unmistakably clear: a system with total context of your work and personal life, gated behind user permissions rather than built-in privacy limits. The usefulness case is real, but so is the discomfort — and going by the public reaction so far, OpenAI has real work to do convincing people the trade-off is worth it.

Sources: 2oceansvibe News, The Comeback