Electronic monitoring (EM) has undergone successful testing and implementation as a fisheries monitoring tool since the early 2000s, and has more recently been expanded through artificial intelligence approaches such as deep learning to automate image analysis. Effective EM implementation can enhance both fishers self-reporting, and on-board observer programs. Advantages of EM over other methods include health and safety improvements associated with not sending observers on vessels, flexible data collection that does not rely on observer availability, and a permanent video record available for review by experts. Image analysis in EM is time consuming and expensive. Implementing machine assisted object detection of fish using deep learning algorithms has the potential to significantly decrease the time necessary to analyse the video imagery. A pilot study is currently being done to test the effectiveness of EM assisted by deep learning algorithms in the identification of species caught in Western Australian shark fisheries. Preliminary results show an accuracy of between 78 and 90% for key species on the test (leave out) dataset. Further analysis is needed as more data is made available and more species are added to the model. This semi-automated approach highlights the potential of artificial intelligence in EM of multispecies fisheries.