Osama Zafar
3 indexed papers
Publications per year
Top categories
Frequent co-authors
Research Timeline
The paper argues that current AI content watermarking benchmarks fail to test for bias across different languages, cultures, and demographics, proposing a new set of evaluation standards to ensure fairness in content provenance.
The paper argues that Agentic AI fundamentally breaks the historical security tradeoff between deception fidelity and scale, necessitating a shift from authenticating actors to evaluating actions.
The paper introduces a Privacy Policy Enforcement (PPE) framework using dual one-class density estimators to detect contextual data leakage in Retrieval-Augmented Generation (RAG) systems, achieving high accuracy with low false positives.
Papers
Privacy Policy Enforcement Guardrails for Data-Sensitive Retrieval-Augmented Generation
Osama Zafar, Alexander Nemecek, Yiqian Zhang, Wenbiao Li +4 more
The paper introduces a Privacy Policy Enforcement (PPE) framework using dual one-class density estimators to detect contextual data leakage in Retrieval-Augmented Generation (RAG) systems, achieving h…