Sam Altman has revealed that OpenAI is working toward an AI assistant that could continuously monitor your computer screen, attend your meetings, and access your emails and documents to provide real-time work suggestions. In a recent interview, the OpenAI CEO estimated this capability could become genuinely useful within the next six months, marking a significant shift in how people interact with artificial intelligence.

What Exactly Could This New ChatGPT Version Do?

Altman described a system that would maintain constant awareness of a user's digital activities and work context. Rather than requiring users to explain their situation each time they ask for help, this AI assistant would already understand what they are working on, who they are communicating with, and what information is relevant to their current task.

The proposed assistant would be able to draw on information from multiple sources while a person is working, offering context-sensitive suggestions without making decisions independently. Users would retain full control over which applications and services the AI can access, such as:

• Screen Monitoring: The AI would watch your computer display to understand what you are currently working on and the context of your tasks.
• Meeting and Call Recording: The system could attend meetings and record calls to build a comprehensive understanding of your work discussions and decisions.
• Information Access: Users could connect the AI to email, text messages, documents, Slack, and other services they choose to share with it.

"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," said Sam Altman, CEO of OpenAI.

Sam Altman, CEO of OpenAI

Altman explained that this capability would be particularly valuable for people managing large amounts of information and competing priorities. He used the example of a startup CEO who cannot possibly read every piece of customer feedback or keep track of all ongoing projects. With this AI assistant, such a person could have the system working alongside them, offering suggestions as they write a sales pitch or strategy document, or even handling smaller tasks to free up their time.

How Is OpenAI Already Moving in This Direction?

OpenAI has already begun implementing features that move toward this vision. The company recently introduced Computer History for the ChatGPT desktop app on Mac, which allows ChatGPT to keep track of activity across selected apps and websites. Rather than relying on screenshots, this feature uses activity data available through macOS accessibility features, including clicks, typing, keyboard shortcuts, and switching between applications.

Users must manually enable Computer History and can decide which apps and websites contribute to the system's understanding of their activities. They can also pause collection or delete stored activity at any time. This current feature is much narrower in scope than what Altman described, but it demonstrates the direction OpenAI is heading.

OpenAI has also already introduced memory features that allow ChatGPT to use information from previous conversations to provide more tailored responses. Users can manage memory and chat history settings, including turning off saved memories entirely.

What About Privacy and Security Concerns?

The development of screen-level AI assistance has raised legitimate security and privacy considerations. Microsoft, which is pursuing a similar approach through its Recall feature for Copilot+ PCs, faced scrutiny when researchers identified ways to access stored snapshot data in early preview builds. Microsoft subsequently added additional controls and security requirements around the feature.

Microsoft says Recall is opt-in, stores snapshots locally, and encrypts them on the hard drive. Users can pause the feature, delete snapshots, and control which content is captured. Altman emphasized that his proposed system would similarly require users to actively decide what information and services the AI could access, rather than granting blanket permission.

When Could This Actually Happen?

When asked by interviewer Cory Levy, founder of Z Fellows, when such a system could become genuinely useful, Altman gave a specific timeline: "Sometime within the next six months". However, it is important to note that this is Altman's estimate of when the technology could become practically useful, not an announced launch date for a specific OpenAI product. He did not name a product, confirm a release date, or lay out a final set of capabilities during the interview.

Altman suggested that the leap to a much more capable system may not be far away. He stated that "we're only one model generation away from this actually being incredibly useful," referring to the next iteration of OpenAI's underlying AI model. This implies that the company believes its next major model release could enable the kind of continuous context awareness and proactive assistance he described.

Altman

The vision Altman outlined represents a fundamental change in how people might work with AI. Instead of treating ChatGPT as a tool you consult when you need help, it would become an always-present assistant that understands your work, your goals, and your constraints, offering suggestions and handling tasks in real time. Whether this vision becomes reality within six months remains to be seen, but OpenAI's current trajectory suggests the company is actively working toward this goal.