Story perspectives

AI-Powered YAM-9 Satellite Identifies Earth Features, Sparks Surveillance Debate

7/2/2026

Summary

Story summary

YAM-9 satellite used AI under NAVI-Orbital program (NASA JPL and Loft Orbital) to identify features from natural language prompts without ground control.
It achieved 88.2 % accuracy on 7,960 images across categories like residential and beach, but only two live in-orbit captures have been made.
Researchers say the system could monitor oil spills or border activity, but they warn of surveillance concerns and will test robustness against adversarial prompts before use.

1 of 1

Story summary

YAM-9 satellite used AI under NAVI-Orbital program (NASA JPL and Loft Orbital) to identify features from natural language prompts without ground control.
It achieved 88.2 % accuracy on 7,960 images across categories like residential and beach, but only two live in-orbit captures have been made.
Researchers say the system could monitor oil spills or border activity, but they warn of surveillance concerns and will test robustness against adversarial prompts before use.