Being critical about magnetic survey practise and how to revisit older surveys, and uplift the information content to modern standards, is an on-going challenge. Many data repositories have 20 + year old surveys that remain the only alternatives, for mounting a fresh campaign to explore under cover. With 50 years of continuous improvements in automation in geophysics, a solid case can be put that ML and AI is applicable. However, the older surveys and their limitations, present some complex requirements for decision making that are not yet automatable. The geology side is still emerging from being a descriptive science. Structural geology rules and an implicit engine based upon geostatistical methods has emerged as the front runner for ML in geology. In the case reported here, creating the Intrepid database, levelling, choosing Trend Gridding to improve the ability to honour underlying gradients and resolution, decorrugation, Reduction to the Pole, Worming, Cauchy Integration, Down wards continuation for depth constraints and plunge, and then looking at a first 3D model, all repeatable, semi-automatic, processes. What used to take months can be done in less than a day. However, the sequence and the checking remain a human only activity. Managing expectations about the role of Artificial Intelligence in exploration geoscience, is tough.

Open-Access Online Publication: May 29, 2023