309: What AI Changes About Seismic Interpretation

“If you keep training on what people have already discovered and what’s already been identified, then you’re gonna be limited in what you can actually come up with.”

AI is very good at finding patterns in seismic data. The harder question is whether those patterns make geological sense.

Nancy House remembers looking at a Gulf of Mexico seismic line after years of working in an overthrust regime and initially seeing a thrust fault. It wasn’t. Her regional experience had shaped what she saw. That is exactly the kind of mistake AI can make if pattern recognition gets separated from geological context.

Nancy connects that problem to two years of work with the GeoGPT governance committee, where the questions are just as unsettled: what knowledge trains these systems, how proprietary data is protected, and who should set the boundaries as the technology keeps changing.

QUESTIONS THIS CONVERSATION ANSWERS

> How can interpreters tell when AI has recognized geology and when it has only recognized a familiar pattern?

> Could training AI on existing interpretations limit the ability to identify genuinely new geologic possibilities?

> What role should professional societies play in governing geoscience AI?

ABOUT SEISMIC SOUNDOFF

Seismic Soundoff showcases conversations addressing the challenges of energy, water, and climate. Produced by the Society of Exploration Geophysicists (SEG) and hosted by Andrew Geary of 51 features, these episodes celebrate and inspire the geophysicists of today and tomorrow. Three new episodes monthly. See the full archive at https://seg.org/resources/podcast/.