In the face of dramatic climate change and human pressure acting on remaining forest areas across tropical, temperate and boreal biomes, there has emerged a coordinated effort to identify and protect forests that are currently considered "intact". These forests are hypothesized to be more resilient to future abiotic perturbations than fragmented or degraded forests, and therefore, will provide more reliable carbon storage and/or biodiversity services into an uncertain future. Research in the fields of contemporary and paleoecology can offer valuable insights to enhance our ability to assess resilience of forests and whether these would be comparable across forest biomes. Contemporary ecological monitoring has been able to capture processes acting over the short-to-medium term, while paleoecological methods allow us to derive insights of the long-term processes affecting forest dynamics. Recent efforts to both identify intact forests, based on area definitions, and assess vegetation climate sensitivity globally have relied on satellite imagery analysis for the time period 2000-2013. In this paper, we compare these published datasets and do find that on average intact forests in boreal and tropical biomes are less sensitive to temperature and water availability, respectively; however, the patterns are less clear within biomes (e.g., across continents). By taking a longer perspective, through paleoecology, we present several studies that show a range of forest responses to past climatic and human disturbance, suggesting that short-term trends may not be reliable predictors of long-term resilience. We highlight that few contemporary and paleoecology studies have considered forest area when assessing resilience and those that have did find that smaller forest areas exhibited greater dynamism in species composition, which could be a proxy for declining resilience. Climatic conditions in the Anthropocene will be pushing forest systems across biomes into novel climates very rapidly and with current knowledge it is difficult to predict how forests will react in the immediate term, which is the most relevant timeframe for global efforts to reduce carbon emissions.