Robust and Calibrated Detection of Authentic Multimedia Content

TL;DR

This paper introduces a resynthesis-based detection framework that reliably verifies multimedia authenticity with low false positives and robustness against resource-efficient adversaries, addressing key limitations of existing deepfake detectors.

Contribution

The paper presents a novel resynthesis approach that improves reliability and robustness in multimedia authenticity verification, especially against computationally efficient adversaries.

Findings

Achieves low false positive rates in authenticity verification.

Demonstrates robustness against resource-constrained adversaries.

Supports multiple modalities with state-of-the-art inversion techniques.

Abstract

Generative models can synthesize highly realistic content, so-called deepfakes, that are already being misused at scale to undermine digital media authenticity. Current deepfake detection methods are unreliable for two reasons: (i) distinguishing inauthentic content post-hoc is often impossible (e.g., with memorized samples), leading to an unbounded false positive rate (FPR); and (ii) detection lacks robustness, as adversaries can adapt to known detectors with near-perfect accuracy using minimal computational resources. To address these limitations, we propose a resynthesis framework to determine if a sample is authentic or if its authenticity can be plausibly denied. We make two key contributions focusing on the high-precision, low-recall setting against efficient (i.e., compute-restricted) adversaries. First, we demonstrate that our calibrated resynthesis method is the most reliable approach for verifying authentic samples while maintaining controllable, low FPRs. Second, we show that our method achieves adversarial robustness against efficient adversaries, whereas prior methods are easily evaded under identical compute budgets. Our approach supports multiple modalities and leverages state-of-the-art inversion techniques.