United States Coast Guard – AI Use Cases

Use Case Name: Adaptive Risk Model for Inspected Small Passenger Vessels 

Use Case ID: DHS-178 

Use Case Summary: This AI system helps the USCG assign a numerical risk score for Small Passenger Vessels (SPV) relative to one another. The SPV Safety Task Force leveraged machine learning (ML) and subject matter expert (SME) input to create an adaptable, quantitative analysis tool that reliably identifies the key drivers of marine casualties and calculates a risk score for each vessel within the largest segment of the U.S.-inspected fleet. A vessel risk score is computed with an analytic model that leverages logistic regression and basic ML. The risk score is then combined with an SME-derived consequence score that considers the number of vessel passengers and operation distance from shore. The final combined number is used to categorize a vessel into one of three tiers: at the lowest risk tier, vessels will receive, at a minimum, mandated inspections as per Title 46 Code of Federal Regulations, while vessels in tiers 2 and 1 may receive more frequent or more thorough inspections by more experienced inspectors. Before an inspection recommendation is communicated to the vessel owner, an SME reviews and approves the vessel scores. Follow-on policy efforts set the required scope and frequency of inspection, the minimum competency of the lead inspector, and the level of accountability for each vessel. In addition to their modeling and policy efforts, this team conducted extensive education and outreach with the field, district, and area commands to aid adoption and foster implementation. Their efforts directly improved marine inspector allocation, targeted risk, and enhanced the governance necessary to increase passenger safety across the country. 

Use Case Topic Area: Mission-Enabling (internal agency support) 

Deployment Status: Deployed (Operation and Maintenance)  

Safety- and/or rights-impacting? No 

Face Recognition/Face Capture (FR/FC)? No 

Use Case Name: Silicon Valley Innovation Program (SVIP) Language Translator 

Use Case ID: DHS-263 

Use Case Summary: USCG personnel often interact with individuals who speak little or no English, making it very challenging to quickly assess and respond to a variety of situations. To better execute these responsibilities, the USCG needs a portable language translator that can rapidly and effectively communicate with non-English speakers. USCG boarding teams need a handheld, universal language translator that works at the speed of natural language and operates in disconnected environments. A handheld voice translator could be used to communicate the questioning of vessels in high-tension situations where a USCG vessel does not have a translator available encounters different or multiple languages. In cases involving a rescue, being able to determine critical information such as the number of people aboard and the status of life-saving equipment is vital to commanders in their decision-making process. Once the USCG boarding team makes contact with the occupants of the boarded vessel, they must be able to communicate and give orders to the occupants who may not speak English or may speak multiple languages. The development of an offline language translator could enable accurate and swift translation of orders and directions, which are critical to the safety and security of the boarding team and the occupants of the vessel. This use case is funded by DHS (S&T), Silicon Valley Innovation Program (SVIP), 

Deployment Status: Inactive (research and development only).  This use case was reported in a previous version of the DHS AI Use Case Inventory but is a research and development use case that is not planned to be deployed.