This article provides a comprehensive review of the critical role of detection efficiency in demonstrating non-classicality across various device-independent and semi-device-independent scenarios. The central focus is the detection loophole , a challenge in which imperfect detectors can allow classical hidden variable models to mimic quantum correlations, thus masking genuine non-classicality. As a review, the article revisits the paradigmatic Bell scenario, detailing the efficiency requirements for the CHSH inequality—such as the 2/3 threshold for symmetric efficiencies—and traces the historical trajectory toward the first loophole-free tests. The analysis extends to other causal structures to explore how efficiency requirements are affected in different contexts. These include: the instrumental scenario, which for binary variables has recently been shown to follow the same inefficiency bounds as the bipartite dichotomic Bell scenario; the prepare-and-measure scenario, where inefficiencies impact the certification of a quantum system’s dimension and create security breaches in protocols such as Quantum Key Distribution (QKD); and the bilocality scenario, which exemplifies how employing multiple independent sources can significantly relax the required efficiencies to certify non-classical correlations.
We introduce the dimension-restricted prepare and broadcast (PAB) scenario, which generalizes standard prepare and measure frameworks. Here, the system prepared by a sender undergoes a broadcasting transformation before being locally measured by multiple receivers. We develop a hierarchy of classical, quantum, and nonsignalling models describing this scenario, characterize their corresponding correlation sets, and derive new families of Bell-like inequalities together with linear and semidefinite programming methods for their certification. First, assuming shared randomness, we prove that the hierarchy collapses into a single set whenever we consider only one measurement per party. Then, considering multiple possible measurements, we show that PAB scenarios can activate nonclassicality of noisy broadcasting resources and can reveal preparation nonclassicality more robustly than large finite families of standard prepare and measure tests.