A significant amount of global fishing is illegal, unreported, or unregulated, equaling roughly a third of all legally caught fish. Illegal fishing is often conducted with prohibited gear and results in overfishing, which damages coral reefs, seamounts, and other vulnerable marine ecosystems. Approximately $55\%$ of the world’s reefs are threatened by excessive and destructive fishing, with nearly $95\%$ of reefs affected in South Asia. Illegal fishing also results in the capture of endangered species, such as the Atlantic salmon, or non-target species like sharks and turtles.

Because the fishing industry is inherently difficult to monitor, organizations are leveraging artificial intelligence (AI) to tackle this problem. For example, The Nature Conservancy launched a challenge on Kaggle to source a machine learning model capable of detecting when protected fish are caught. The AI software flags suspicious portions of fishing video footage for human review. Given that it currently takes $6$ hours for an officer to manually review $10$ hours of fishing footage, this AI technology significantly increases the efficiency of monitoring and the scrutiny placed on illegal fishing operations.