Entity · paper

Phantoms and Disclosures: a Causal Framework for Auditing Synthetic Data

paperactivephantoms-and-disclosures-a-causal-framework-for-auditing-synthetic-data-8a9e2018·1 events·first seen Jun 16, 2026

Aliases: Phantoms and Disclosures: a Causal Framework for Auditing Synthetic Data

Co-occurring entities

Differential Privacy

More like this (12)

Beyond Third-Person Audits: Situated Interaction Auditing for User-Centered LLM Bias Research Reason-Mediated Behavioral Models for Auditing LLM Social Simulators synthetic data evaluation Valid Inference with Synthetic Data via Task Exchangeability Bayesian Inference and Decision Audits for Public Archives of Frontier AI Evaluations Relaxing Faithfulness with Intervention-Only Causal Discovery Dynamic Capability Scoping for Enterprise AI Agents: A Synthetic Dataset and Three-Source Permission Architecture From Observation to Intervention: A Causal Audit of Expert Importance in Mixture-of-Experts Models Artificial Analysis Conversational Dynamics Does Bielik Know What It Doesn't Know? Activation Dispersion Separates Entity Familiarity from Factual Reliability Across Model Scale Evaluating RE Practices for Explainability: Synthesizing Insights from Daimler Truck into an Explainable RE Framework Proposal The Test Oracle Problem in Synthetic LLM-as-Judge Corpora: Disappearance, Distortion and a Validation Protocol

Recent events (1)

6arXiv · cs.AI·Jun 16, 2026·source ↗

Causal auditing framework detects privacy disclosures in synthetic data without model access

A new arXiv preprint introduces a model-agnostic empirical framework for auditing synthetic data generated by LLMs and generative AI systems for privacy leakage. The framework distinguishes 'true disclosures' (direct reproduction of user data) from 'phantom disclosures' (incidental generation), using held-out control sets and statistical hypothesis testing without requiring model access, canary insertion, or shadow model training. It functions as a membership inference attack and provides empirical lower bounds on privacy leakage that are tighter than prior data-based auditing methods. The approach is computationally lightweight and applicable to any synthetic data generation mechanism.

Evaluation and Benchmarking AI Safety Research Differential Privacy Phantoms and Disclosures: a Causal Framework for Auditing Synthetic Data