Artificial intelligence and machine learning are two much-heralded technologies set to revolutionise signals intelligence collection, processing and dissemination.

An increasingly congested radio spectrum is set to challenge the limits of human cognition in the search for the signal of interest. In this episode of Radioflash! we catch up with Patrick ‘Krown’ Killingsworth, EpiSci’s  director of autonomy projects.

We define the terms Artificial Intelligence (AI) and Machine Learning (ML) and their unique capabilities. ML algorithms are trained on huge amounts of data which makes these algorithms great for processing huge amounts of data, a key requirement in Signals Intelligence (SIGINT). The application of AI and ML in SIGINT is not necessarily about replacing the human but helping the SIGINT operator sort these data. Nonetheless, the continuing introduction of AI and ML into SIGINT analysis prompts concern and enthusiasm in equal measure. Ensuring that enough data are available for training algorithms creates challenges given the paucity of data which the SIGINT cadre may be interested in.

We talk about the risks of using synthetic data for training and tackling the risk by anticipating potential problems from the start. AI- and ML-enabled SIGINT systems continue to get smarter, faster and more accurate. The future brings challenges in terms of moving this SIGINT technology from the strategic level to the tactical edge, although edge computing in the tactical domain should help no end in this regard.