To estimate human exposure to methylmercury (MeHg), risk assessors often assume 95%–100% bioavailability in their models. However, recent research suggests that assuming all, or most, of the ingested mercury (Hg) is absorbed into systemic circulation may be erroneous. The objective of this paper is to review and discuss the available state of knowledge concerning the assimilation or bioavailability of Hg in fish and humans. In fish, this meant reviewing studies on assimilation efficiency, that is the difference between ingested and excreted Hg over a given period of time. In humans, this meant reviewing studies that mostly investigated bioaccessibility (digestive processes) rather than bioavailability (cumulative digestive + absorptive processes), although studies incorporating absorption for a fuller picture of bioavailability were also included where possible. The outcome of this review shows that in a variety of organisms and experimental models that Hg bioavailability and assimilation is less than 100%. Specifically, 25 studies on fish were reviewed, and assimilation efficiencies ranged from 10% to 100% for MeHg and from 2% to 51% for Hg(II). For humans, 20 studies were reviewed with bioaccessibility estimates ranging from 2% to 100% for MeHg and 0.2% to 94% for Hg(II). The overall absorption estimates ranged from 12% to 79% for MeHg and 49% to 69% for Hg(II), and were consistently less than 100%. For both fish and humans, a number of cases are discussed in which factors (e.g., Hg source, cooking methods, nutrients) are shown to affect Hg bioavailability. The summaries presented here challenge a widely-held assumption in the Hg risk assessment field, and the paper discusses possible ways forward for the field.