OpenAI CEO Sam Altman revealed the company’s future plans for its artificial intelligence model, ChatGPT. The next version, GPT-6, will be more personalized and adaptable than previous models, allowing users to define tone and personality. This feature is expected to make the AI system more human-like.
Altman emphasized the importance of memory in creating a truly personal experience with ChatGPT. He wants the model to remember individual preferences, routines, and quirks to adapt accordingly. To achieve this, OpenAI has been working closely with psychologists to understand user behavior and well-being.
The company’s commitment to ideological neutrality and customizability was also highlighted by Altman. GPT-6 will be designed to provide a neutral stance while allowing users to push it further in their preferred direction.
Following the rocky rollout of GPT-5, Altman acknowledged the need for improvement and confirmed that a tone update had been quietly pushed to the model, making it warmer and more helpful.
Altman also addressed concerns over privacy, particularly regarding temporary memory not being encrypted. He confirmed that encryption will be added in the future, ensuring sensitive information remains protected.
As OpenAI looks to expand its product range, Altman mentioned his interest in brain-computer interfaces and other areas such as energy, novel substrates, robots, and faster data center construction. However, for now, ChatGPT remains the core consumer product, with Altman focusing on making it more flexible and useful in daily life.
Despite the potential benefits of further AI development, Altman warned that there are limits to the chat use case, suggesting that models may not get much better or even worse over time.
Source: https://www.cnbc.com/2025/08/19/sam-altman-on-gpt-6-people-want-memory.html