As autonomous shipping advances, reliable and efficient voice communication remains a challenge. This presentation will showcase a domain-specific multilingual maritime speech recognizer that has been enhancing maritime operations by transcribing and processing VHF radio communications. Expanding on this, a new concept for autonomous vessels has been designed to enable AI-driven voice interaction, supporting seamless ship-to-ship and ship-to-shore communication. By integrating advanced maritime speech recognition with technologies like text-to-speech and natural language processing, this innovation has the potential to create a more connected and responsive communication framework between autonomous and crewed vessels, enhancing both safety and efficiency.