OpenAI CEO Sam Altman has put a strikingly short timeline on a future where ChatGPT could understand far more than what users type into a chat box.
In a conversation with Cory Levy, founder of Z Fellows, Altman said a future “descendant of ChatGPT” could watch a user’s computer screen, meetings and calls, while drawing context from services such as email, texts, documents and Slack. He suggested such an assistant could become useful within the next six months.
“I think we are close to a world where you can have a descendant of ChatGPT watch your computer screen all the time, watch every meeting you’re in, record every call, everything that has perfect context of your whole life, everything you see,” Sam said.
The timeline is Altman’s prediction, not an OpenAI product announcement. But the comments point to a broader shift already underway in AI: moving from responding to prompts towards understanding the user’s wider digital context.
Sam Altman Envisions Persistent Context
During a conversation with Z Fellows founder Cory Levy, Altman described a future assistant that could watch a user’s computer screen and meetings, record calls and build what he called “perfect context” of a person’s digital life. He also said users would control which information and services the system could access.
Altman gave a workplace example involving a startup chief executive who cannot personally keep up with every customer response or discussion.
With access to the relevant information, he suggested, an AI assistant could provide suggestions while the person is writing a sales pitch or strategy document, flag a possible mistake or offer to perform a task.
Importantly, these are examples of what Altman believes future AI assistants could do. They should not be read as confirmation that ChatGPT currently has continuous access to a user’s screen, calls or meetings.
ChatGPT Already Sees Screens
OpenAI has already introduced some of the building blocks behind this vision.
Eligible subscribers can currently share their screen or live video with ChatGPT during Advanced Voice conversations on iOS and Android. Users actively initiate and stop screen sharing through the app or their device controls.
However, this is materially different from the always-on system Altman described. OpenAI’s latest Live voice experience does not currently support video or screen sharing, while those capabilities remain available through Advanced Voice for eligible subscribers.
ChatGPT has also expanded its ability to work with information outside an individual conversation. OpenAI supports connected apps that can bring information from external services into ChatGPT, while access and permissions can vary by plan and workspace.
Memory Is Moving Forward
Another part of the shift is memory.
In June 2026, OpenAI introduced a more capable memory architecture called “Dreaming”, designed to synthesise information from conversations and keep ChatGPT’s understanding of a user’s context fresher and more relevant over time. OpenAI described the system as an evolution of its earlier memory capabilities rather than simply a larger list of saved facts.
That matters because an assistant capable of seeing screens or meetings would become significantly more useful when combined with persistent context. Instead of treating every interaction as an isolated request, the system could potentially connect information across different tasks.
But that same capability would increase the amount and sensitivity of information an AI could potentially process.
Privacy Could Determine Adoption
An assistant that can access emails, documents, meetings and screens would encounter considerably more sensitive information than a conventional chatbot. Corporate presentations, customer data, private conversations and confidential documents could all appear within the assistant’s potential field of view.
OpenAI already provides controls for some existing forms of data. Temporary Chats do not appear in chat history, do not create memories and are not used to improve OpenAI’s models. OpenAI says it may retain a copy of Temporary Chats for up to 30 days for safety purposes.
The challenge for any future screen- or call-aware assistant would be more granular: users would need to understand exactly what is being captured, when it is being processed, what is retained and which applications or people are included.
Recall Shows The Challenge
Microsoft’s Recall offers a useful comparison, although it is not identical to Altman’s proposed system.
Recall can periodically save snapshots of activity on compatible Windows PCs so users can later search what they have seen.
Microsoft says saving snapshots is off by default and requires users to opt in. It also says the snapshots and related information are stored locally and encrypted, while users can pause, delete or disable them. Recall does not record audio or save continuous video.
The comparison illustrates why the next stage of AI competition will not be determined solely by model intelligence. Permission systems, transparency and data controls could become equally important as assistants gain access to more of a user’s digital environment.
ChatGPT has expanded memory, voice, screen sharing and connected-app capabilities. If those technologies converge, the defining question for the next generation of AI assistants may be not simply what they can see, but how much users are willing to let them see.
Also Read: PM CARES Donations Fell 30%, So Why Did Its Corpus Rise To ₹8,452 Crore?
