While the fully mature commercial applications of deeply fused "Agentic AI personality companions" are still in their early stages, leading universities, research labs, and forward-thinking researchers are already actively exploring and deploying hybrid frameworks that blend MBTI with Enneagram (or similar psychological models) to power more human-like AI agents. Key examples include:ETH Zurich: In their 2025 paper "Psychologically Enhanced AI Agents", researchers introduced the MBTI-in-Thoughts (MiT) framework, which infuses psychological depth directly into agent reasoning. University of Tokyo: Studies demonstrate that LLM agents in multi-agent systems naturally evolve distinct MBTI-like personalities through social interactions—shifting from INFJ-style patterns toward ENTJ behaviors, for instance—creating genuine individuality and emotional synchronization. Other notable work is underway at Stanford, Google DeepMind, and similar institutions, focused on building AI agents that authentically simulate real human personality traits and dynamics.

C@C

(view original post →)