Allosteric communication is established by networks through which strain energy generated at the allosteric site by an allosteric event, such as ligand binding, can propagate to the functional site. Exerted on multiple molecules in the cell, it can wield a biased function. Here, we discuss allosteric networks and allosteric signaling bias. Networks are graphs specified by nodes (residues) and edges (their connections). Allosteric bias is a property of a population. It is described by allosteric effector-specific dynamic distributions of conformational ensembles, as classically exemplified by G protein-coupled receptors (GPCRs). An ensemble describes the likelihood of a specific (strong/weak) allosteric signal propagating to a specific functional site. A network description provides the propagation route in a specific conformation, pinpointing key residues whose mutations could promote drug resistance. Efficiency is influenced by path length, relative stabilities and allosteric transitions. Through specific contacts, specific ligands can bias signaling in proteins, for example, in receptor tyrosine kinases (RTKs) toward specific phosphorylation sites and cell signaling activation. Thus, rather than the two – active and inactive – states, and a single pathway, we consider multiple states and favored pathways. This allows us to consider biased allosteric switches among minor, invisible states and observable outcomes. Within this framework, we further consider signaling strength and duration as key determinants of cell fate: If weak and sustained, it may induce differentiation; If bursts of strong and short, proliferation.