AI and Machine Learning in Mirikat Blessings

 

This project harnesses AI and machine learning to reimagine and increase access to the ancient Ethiopian tradition of Mirikat, the sacred practice of offering blessings. By integrating natural language processing (NLP) models and retrieval-augmented generation (RAG) techniques, the system interprets user-submitted concerns and generates deeply spiritual, poetic blessings in response. Currently, the AI-driven backend utilizes language models to craft personalized blessings, ensuring accessibility to users worldwide.

 

We will continue to work on future iterations expanding upon this foundation by incorporating multilingual support, beginning with Amharic, to preserve and honor the cultural depth of mirikat in its native form. Additionally, our intentions are to develop enhancements through fine-tuned large language models (LLMs) and advanced retrieval mechanisms to allow for more context- aware blessings, drawing from historical texts, oral traditions, and user interactions.

 

Our hope is that by making this sacred ritual accessible through AI, our project interweaves art-making, ancient wisdom, and modern technology, offering blessings to anyone, anywhere. Our work seeks to preserve mirikats' traditional cultural heritage while simultaneously evolving its reach through machine learning and computational creativity.

​

User-authored inputs are not saved or visible on the backend.
The model is not being trained or updated based on anything the users input.
No data is shared with or used by any commercial LLM.
Everything is private and ephmeral.

​