Since 2020, the ‘Antarctic Soundscapes’ (AS) project has aimed to study anthropogenic impact in Antarctica through sound by deploying a network of AudioMoth recorders at key locations across the Fildes Peninsula (King George Island). These devices continuously capture bioacoustic data during summer campaigns, providing insights into the impacts of human activity on this ecosystem. However, the project faces two significant challenges: the limited autonomy of the recording devices, which typically require battery replacements every 15-20 days, and the extensive human effort needed to process and analyze months of continuous recordings. To address these limitations, we enhanced the AudioMoth-based recording system by integrating a solar panel, rechargeable battery, and charge controller, enabling continuous operation for up to 84 days without human intervention during the 2022-2023 campaign. Additionally, we implemented a machine learning-based audio tagging system that automates the identification of anthropogenic sounds, reducing the time and resources required for manual analysis. These contributions represent a significant step towards fully autonomous monitoring and analysis of Antarctic soundscapes, enabling more efficient and sustainable research.