GOD model: Privacy Preserved AI School for Personal Assistant

TL;DR

The GOD framework tackles privacy concerns in proactive on-device personal AI by running evaluation and training inside a Trusted Execution Environment (TEE). It introduces curriculum-based exams that simulate realistic user needs to address cold-start and autonomy while preserving privacy via Data Connectors and on-device Personal AIs. Key contributions include a TEE-based secure evaluation, curriculum-based assessment, data-value estimation, and anti-gaming safeguards, all enabled by mechanisms like a data flywheel and token incentives. This approach enables scalable, privacy-preserving personalization with transparent verification and continuous improvement for trusted on-device assistants $\Total Score = w_C \cdot Coverage + w_Q \cdot Quality + w_F \cdot Freshness$.

Abstract

Personal AI assistants (e.g., Apple Intelligence, Meta AI) offer proactive recommendations that simplify everyday tasks, but their reliance on sensitive user data raises concerns about privacy and trust. To address these challenges, we introduce the Guardian of Data (GOD), a secure, privacy-preserving framework for training and evaluating AI assistants directly on-device. Unlike traditional benchmarks, the GOD model measures how well assistants can anticipate user needs-such as suggesting gifts-while protecting user data and autonomy. Functioning like an AI school, it addresses the cold start problem by simulating user queries and employing a curriculum-based approach to refine the performance of each assistant. Running within a Trusted Execution Environment (TEE), it safeguards user data while applying reinforcement and imitation learning to refine AI recommendations. A token-based incentive system encourages users to share data securely, creating a data flywheel that drives continuous improvement. Specifically, users mine with their data, and the mining rate is determined by GOD's evaluation of how well their AI assistant understands them across categories such as shopping, social interactions, productivity, trading, and Web3. By integrating privacy, personalization, and trust, the GOD model provides a scalable, responsible path for advancing personal AI assistants. For community collaboration, part of the framework is open-sourced at https://github.com/PIN-AI/God-Model.

Figures (6)

• Figure 1: The GOD model, running in a secure TEE, generates realistic user queries. The Personal AI on the user’s device processes private data to answer the queries. The GOD model verifies and grades those answers through internal checks or external APIs while ensuring privacy.
• Figure 2: The GOD model evaluates the Personal AI by posing personal questions (e.g., height, last email subject, recent online orders). The Personal AI uses only private, on-device data to respond without revealing raw information.
• Figure 3: The GOD model uses Data Connectors to securely access user data within a TEE, then scores the Personal AI’s responses in multiple categories.
• Figure 4: Level 1 (Easy): Direct Factual Queries
• Figure 5: Level 2 (Medium): Cross-Referencing and Context