In the current Anthropocene era, humans are having an increasingly marked impact on wildlife populations and ecosystems. This “super-predator” now exerts the most influential selective force shaping the evolution of the behaviour of wild animals. While we are witnessing the return of large carnivores in many European countries, human activities undoubtedly alter the interactions between these predators and their prey, questioning the role that carnivores play in shaping ecosystem functioning in human-dominated landscapes. In this context, the AnthropoFEAR project will evaluate how the predation context (large carnivores, hunting, both) influences the behavioural tactics of roe deer (a common prey for both lynx and wolf, and an intensely hunted herbivore). The project will employ both observational (based on GPS and behavioural monitoring in several roe deer populations of Europe) and experimental approaches. We will first compare anti-predator tactics (in terms of habitat selection, vigilance levels and escape decisions) of roe deer in response to risk in contrasted predation contexts. We will then quantify their cascading effects on the understorey vegetation in a highly controlled ecosystem, in which we will experimentally modify the landscape of fear perceived by animals. Because individuals do not respond in the same way to a given predation risk, we will also assess the degree of inter-individual variability of these anti-predator tactics within populations. These results will reveal the potential role of human activities for driving impoverishment in behavioural variation in hunted large herbivores, a critical component for the resilience of wild populations to environmental modifications linked to global change. Therefore, by combining new data and original approaches based on the intensive monitoring of individuals in long-term studies of large herbivores across Europe, AnthropoFEAR will provide an integrative view of the potential consequences of the return of large carnivores for their prey in anthropogenic ecosystems.