A Responsible Generative Artificial Intelligence Based Multi-Agent Framework for Preserving Data Utility and Privacy
Bibliographic record
Abstract
The exponential growth in the usage of textual data across industries and data sharing across institutions underscores the critical need for frameworks that effectively balance data utility and privacy. This paper proposes an innovative agentic AI-based framework specifically tailored for textual data, integrating user-driven qualitative inputs, differential privacy, and generative AI methodologies. The framework comprises four interlinked topics: (1) A novel quantitative approach that translates qualitative user inputs, such as textual completeness, relevance, or coherence, into precise, context-aware utility thresholds through semantic embedding and adaptive metric mapping. (2) A differential privacy-driven mechanism optimizing text embedding perturbations, dynamically balancing semantic fidelity against rigorous privacy constraints. (3) An advanced generative AI approach to synthesize and augment textual datasets, preserving semantic coherence while minimizing sensitive information leakage. (4) An adaptable dataset-dependent optimization system that autonomously profiles textual datasets, selects dataset-specific privacy strategies (e.g., anonymization, paraphrasing), and adapts in real-time to evolving privacy and utility requirements. Each topic is operationalized via specialized agentic modules with explicit mathematical formulations and inter-agent coordination, establishing a robust and adaptive solution for modern textual data challenges.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.003 | 0.007 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.003 |
| Scholarly communication | 0.003 | 0.003 |
| Open science | 0.002 | 0.003 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.002 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".