AI hallucinations are one of the biggest reasons the mining industry continues to resist adopting AI workflows. But the discussion is often framed incorrectly.
Engineering deliverables already pass through multiple review layers because humans are not perfect either — discipline reviews, interdisciplinary reviews, project management reviews, client reviews, and QA/QC processes all exist precisely because human error is a given.
So the real question is not whether AI can make mistakes. It is whether AI can become reliable enough — used correctly — to materially improve mining workflows.
What this video covers
- AI hallucinations vs human engineering errors
- Why QA/QC systems already exist in mining
- How prompting dramatically impacts AI reliability
- Why most professionals still use AI incorrectly
- The rapid pace of frontier model improvement
- Why mining companies risk falling behind by avoiding AI entirely
The Hallucination Control Playbook
Ten prompting rules — plus a production-ready prompt block — you can start using today to make AI outputs survive technical review.
Read the playbookFrancisca Lombard is the founder of LOMexcel — mining consulting and AI advisory for teams that need AI outputs to survive technical review the first time, not the third.