At the 2026 Internapalooza event, OpenAI CEO Sam Altman said that within the next six months, OpenAI’s upcoming AI models will be able to “perfectly understand users’ context.”
Note: Internapalooza is a signature annual networking and speaking event in Silicon Valley and the San Francisco Bay Area designed specifically for interns at technology companies.
In a conversation with Z Fellow founder Cory Levy, Altman said that within the next six months, AI will evolve to “read” users’ screens, record meetings and calls, and gain a comprehensive understanding of their lives.
Altman outlined a future scenario for AI and said that it is not far from becoming reality:
I think we’re very close to a world where users can use future iterations of ChatGPT to watch their screens, record every meeting they attend and every phone call they make, and understand “everything they see,” enabling AI to perfectly understand every aspect of their lives.
Altman said users could allow AI to take over services such as text messages, email, documents, or Slack, while still choosing the scope of AI control and defining what information it can access. This type of AI would not make decisions directly on users’ behalf, but would instead serve primarily as an assistant, working alongside users to help them handle the enormous amount of information involved in their daily work.
Altman gave the example of a startup CEO who may not have enough time to read every item of customer feedback or track every detail. In such situations, AI could help write customer-facing sales pitches and prepare strategy documents.
He also said AI could offer new ideas like a helpful colleague or point out mistakes users might make. For example, AI might remind a user, “There’s another idea here,” or say, “I think you may be making a mistake here; you should consider this.”
Altman said that it may take only one more generation of models for this capability to become “extremely useful” in practice.