LEGG turns drone imagery into ground-level quake damage

This is the kind of AI that matters in disaster response: not a gimmick, but a plausible shortcut from aerial reconnaissance to the ground-level judgment crews need under time pressure.

• –LEGG's real innovation is view translation, giving responders a street-level lens when access is blocked or too dangerous.
• –The paper is still a research prototype, so the win is feasibility and data efficiency, not a ready-to-deploy emergency system.
• –Its strongest value is rapid pre-assessment for dense urban zones, especially where manual inspections would take days or weeks.
• –The model's usefulness will hinge on data quality, geographic transferability, and how well humans can distinguish synthetic cues from real damage.