Project Title: AI-tool for the audio analysis of Maritime Accident Investigations Boards

Description: 

When transportation accidents occur, investigators collect extensive audio recordings – including interviews and communication logs – to understand the sequence of events. Reviewing and analyzing this data is essential but often challenging due to the volume, complexity, and the need for strict confidentiality. 

In this project, EHUB, Rowbotham Invest, and the Danish Maritime Accident Investigation Board (DMAIB) collaborated to develop a secure AI-based tool that helps investigators work more efficiently with audio material. The result is a dedicated user interface that supports transcription, tagging, commenting, and playback – all tailored to the investigative workflow. 

The system runs on infrastructure fully controlled by the project partners, including a physical deployment setup. This ensures high levels of data sovereignty and forms the basis for future offline (“air-gapped”) solutions. 

By combining investigative insight with technical expertise in AI, software, and hardware, the project paves the way for more effective accident investigations. 

The solution also points to wider possibilities: in any domain where sensitive audio needs to be transcribed and analyzed securely – such as defense, health, or legal – the technology and deployment model developed here could be adapted to meet strict data control requirements. 

Duration: 

Start: 1 April 2025 

End: 30 June 2025