AI shakes up geology with a rock solid machine learning approach.

  • AI and machine learning are being utilized in the field of geology to explore and understand complex geological provinces, such as the western Gawler Craton in Australia.
  • Machine learning can serve as an unbiased check against human data interpretation, providing more objective results.
  • Geologists emphasize the need for human expertise in interpreting the results generated by AI and machine learning, cautioning against over-reliance on technology.

AI and machine learning technologies are revolutionizing the field of geology, with organizations like the Geological Society of South Australia (GSSA) utilizing these technologies to explore and understand complex geological regions. The GSSA has employed these techniques to study the western Gawler Craton, an ancient geological province that houses a variety of critical minerals.

Senior geologist Dr. Mark Pawley emphasized the value of machine learning as a tool for “lineament mapping”, the process of examining linear patterns in a landscape that suggest geological features. According to Dr. Pawley, the unbiased nature of machine learning algorithms offers a helpful counterbalance to human interpretations, which may be influenced by the desire for neat, unbroken structures.

Despite the advantages of machine learning, Dr. Pawley stressed the continued need for human experts in the field of geology. Different rocks carry different levels of magnetism, and interpreting the cause behind a change from magnetically high to magnetically low regions requires human expertise and understanding that current machine learning algorithms do not possess.

This perspective was echoed by Dr. Yusen Ley Cooper, a senior geophysicist with Geoscience Australia, who has worked with machine learning and airborne electromagnetic surveys. He warned against over-reliance on technology for decision-making and emphasized the need for careful supervision in interpreting machine learning output.

Similarly, Dr. Ian Roach, program leader of stratigraphic drilling at Geoscience Australia, noted that while machine learning presents an undeniable advantage in processing large data sets, it is essential to ensure the technology is “well trained”, paralleling it to a well-trained dog.

Overall, while AI and machine learning hold significant potential for the field of geology, experts caution against viewing it as a magic bullet and stress the continued importance of human expertise and oversight.