Despite the importance of visual observation in the ocean, we have imaged a minuscule fraction of the deep seafloor. Sixty-six percent of the entire planet is deep ocean (≥200 m), and our data show that we have visually observed less than 0.001%, a total area approximately a tenth of the size of Belgium. Data gathered from approximately 44,000 deep-sea dives indicate that we have also seen an incredibly biased sample. Sixty-five percent of all in situ visual seafloor observations in our dataset were within 200 nm of only three countries: the United States, Japan, and New Zealand. Ninety-seven percent of all dives we compiled have been conducted by just five countries: the United States, Japan, New Zealand, France, and Germany. This small and biased sample is problematic when attempting to characterize, understand, and manage a global ocean.
Ocean scientists studying diverse organisms and phenomena increasingly rely on imaging devices for their research. These scientists have many tools to collect their data, but few resources for automated analysis. In this paper, we report on discussions with diverse stakeholders to identify community needs and develop a set of functional requirements for the ongoing development of ocean science-specific analysis tools. We conducted 36 in-depth interviews with individuals working in the Blue Economy space, revealing four central issues inhibiting the development of effective imaging analysis monitoring tools for marine science. We also identified twelve user archetypes that will engage with these services. Additionally, we held a workshop with 246 participants from 35 countries centered around FathomNet, a web-based open-source annotated image database for marine research. Findings from these discussions are being used to define the feature set and interface design of Ocean Vision AI, a suite of tools and services to advance observational capabilities of life in the ocean.
The Global Deep Sea Exploration Goals strategy is a spatially balanced, probability-based, actionable global sampling design identifying 10,000 target locations for deep-sea visual observation (≥200 m). This sampling approach integrates four seafloor characteristics: bathymetry, geomorphology, sediment composition, and particulate organic carbon flux, while accounting for documented historical deep-submergence deployments. It aims to correct for historical observational biases across factors such as depth zones, ocean basins, geomorphology, and maritime jurisdictions. These proposed targets will nearly double the number of unique observed seafloor locations, establishing a more geographically and environmentally representative dataset. This globally distributed sampling design also highlights critical scientific gaps in underexplored regions, guiding both exploration and capacity-building priorities. By aligning with and amplifying complementary global initiatives in mapping and sampling, this effort lays the foundation for a more inclusive, coordinated, and statistically representative approach to deep-ocean science, which is essential for informed policy and sustainable resource management.