The goal of this paper is to review current impacts of human activities on the deep-sea floor ecosystem, and to predict anthropogenic changes to this ecosystem by the year 2025. The deep-sea floor ecosystem is one of the largest on the planet, covering roughly 60% of the Earth's solid surface. Despite this vast size, our knowledge of the deep sea is poor relative to other marine ecosystems, and future human threats are difficult to predict. Low productivity, low physical energy, low biological rates, and the vastness of the soft-sediment deep sea create an unusual suite of conservation challenges relative to shallow water. The numerous, but widely spaced, island habitats of the deep ocean (for example seamounts, hydrothermal vents and submarine canyons) differ from typical deep-sea soft sediments in substrate type (hard) and levels of productivity (often high); these habitats will respond differently to anthropogenic impacts and climate change. The principal human threats to the deep sea are the disposal of wastes (structures, radioactive wastes, munitions and carbon dioxide), deep-sea fishing, oil and gas extraction, marine mineral extraction, and climate change. Current international regulations prohibit deep-sea dumping of structures, radioactive waste and munitions. Future disposal activities that could be significant by 2025 include deep-sea carbon-dioxide sequestration, sewage-sludge emplacement and dredge-spoil disposal. As fish stocks dwindle in the upper ocean, deep-sea fisheries are increasingly targeted. Most (perhaps all) of these deep-sea fisheries are not sustainable in the long term given current management practices; deep-sea fish are long-lived, slow growing and very slow to recruit in the face of sustained fishing pressure. Oil and gas exploitation has begun, and will continue, in deep water, creating significant localized impacts resulting mainly from accumulation of contaminated drill cuttings. Marine mineral extraction, in particular manganese nodule mining, represents one of the most significant conservation challenges in the deep sea. The vast spatial scales planned for nodule mining dwarf other potential direct human impacts. Nodule-mining disturbance will likely affect tens to hundreds of thousands of square kilometres with ecosystem recovery requiring many decades to millions of years (for nodule regrowth). Limited knowledge of the taxonomy, species structure, biogeography and basic natural history of deep-sea animals prevents accurate assessment of the risk of species extinctions from large-scale mining. While there are close linkages between benthic, pelagic and climatic processes, it is difficult to predict the impact of climate change on deep-sea benthic ecosystems; it is certain, however, that changes in primary production in surface waters will alter the standing stocks in the food-limited, deep-sea benthic. Long time-series studies from the abyssal North Pacific and North Atlantic suggest that even seemingly stable deep-sea ecosystems may exhibit change in key ecological parameters on decadal time scales. The causes of these decadal changes remain enigmatic. Compared to the rest of the planet, the bulk of the deep sea will probably remain relatively unimpacted by human activities and climate change in the year 2025. However, increased pressure on terrestrial resources will certainly lead to an expansion of direct human activities in the deep sea, and to direct and indirect environmental impacts. Because so little is known about this remote environment, the deep-sea ecosystem may well be substantially modified before its natural state is fully understood.