OpenAI chief Sam Altman says AI assistants could soon maintain continuous context across a user’s digital life, from computer screens and meetings to email and Slack. He believes such a system could become genuinely useful within six months, although OpenAI has not announced a product or launch date matching his prediction.
OpenAI CEO Sam Altman believes the next major step for AI assistants could be a system that understands what users are doing across their digital lives, rather than simply responding to individual prompts.
Speaking to Z Fellows founder Cory Levy, Altman suggested that a future version of ChatGPT could monitor a user’s computer screen, follow meetings and calls, and draw information from services such as email, documents and Slack. He estimated that such an assistant could become useful within the next six months.
The comments point to a shift from chat-based AI towards assistants that retain a much broader understanding of a person’s work. Altman said users would determine which information and services the system could access, with the assistant then using that context to offer suggestions or help complete tasks.
Altman’s vision for an always-on AI assistant
Rather than handing over control of important decisions, Altman described an assistant that would work alongside a user. It could, for example, understand the context behind a document being written or a customer conversation and flag relevant information that the person might otherwise overlook.
“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,” Altman said.
More from Tech
He suggested that the system could be particularly useful for people dealing with large volumes of information. A startup chief executive, for instance, may not have time to read every customer response or keep track of every discussion. An assistant with access to those sources could surface relevant feedback while the user is working.
Altman said the technology would not necessarily act independently. Instead, it could suggest another approach, identify a possible error in a strategy document or offer to carry out a task based on information it already has.
He believes the technology may only be one model generation away from reaching that level of usefulness.
Screen-aware AI already exists, but raises privacy questions
The idea is not entirely new. Microsoft’s Recall feature for Copilot+ PCs already takes snapshots of activity on a Windows computer to help users find previously viewed information. Microsoft says Recall is optional, stores snapshots locally and encrypts them, while also providing controls to pause the feature, delete snapshots and restrict what is captured.
Recall also illustrates the security challenges involved in giving AI access to a detailed record of a person’s digital activity. Earlier preview versions attracted scrutiny after researchers identified potential ways to access stored snapshot information, prompting Microsoft to introduce additional security measures and controls.
OpenAI already offers memory features that can use information from previous interactions to personalise responses, although these do not amount to the continuous, cross-service monitoring Altman described.
Importantly, Altman’s six-month estimate is not a confirmed OpenAI product roadmap. He did not announce a product name, release date or final specification for such an assistant. Any eventual system would also depend on how OpenAI handles permissions, privacy and security when users connect sensitive personal and workplace information.
For now, Altman’s comments offer a glimpse of where he expects AI assistants to head next: away from isolated conversations and towards systems that can understand what a person is doing throughout the day.