This invention describes a system where a user selects an AI virtual assistant that has a distinct personality. This personality influences how the assistant makes decisions by setting limits on its underlying AI algorithms. The system tracks the assistant's performance, adjusts its internal settings to improve future actions, and specifically checks if those adjustments still align with the assistant's original personality, potentially modifying the personality if they don't.
Why it matters: Filed before the widespread adoption of advanced generative AI. Large language models now provide a more robust and dynamic foundation for creating and maintaining consistent personas, making the "quantifiable traits" and "influencing algorithms" aspects of this virtual assistant more feasible and sophisticated.
AI gives you a few directions you could take this. Pick one, and we check whether your version is different enough to patent, then write the filing.
Reinvent this with AI