Chinese researchers have created a system named MemOS, designed to provide artificial intelligence with a persistent, human-like memory. According to a report by Michael Nuñez in VentureBeat, the technology addresses a fundamental limitation that causes AI models to forget information between user interactions.
Current AI assistants often cannot recall past conversations, a problem the researchers call the “memory silo”. This means a user’s preferences or previously discussed details are lost when a new session begins. One notable exception is ChatGPT’s extended memory feature.
The new system treats memory as a core resource, similar to how a computer’s operating system manages processing power. The researchers state that this approach enables AI to genuinely learn and evolve from experience.
In their scientific paper, the team reports that MemOS significantly outperforms existing methods. It showed a 159% performance boost in temporal reasoning tasks compared to OpenAI’s memory systems. This architecture allows AI memories to be portable across different applications and devices, preventing user context from being trapped within a single platform.
The researchers also envision a future marketplace for what they call “paid memory modules”. This would allow domain experts to package their specialized knowledge into units that could be installed into AI systems. The team has released the code as an open-source project to encourage widespread adoption.