Story perspectives

AI Uncovers Ancient Life Signs in 3.3 Billion-Year-Old Rocks

11/21/2025

Summary

Story summary

Researchers developed a machine-learning technique to identify signs of ancient life in rocks up to 3.3 billion years old.
The study published on November 17 shows the AI system distinguished biotic from abiotic samples with over 90% accuracy.
Its confidence decreases for older samples.
The method detected biogenic molecules in 3.3-billion-year-old rocks from South Africa.
Experts say the technique could aid on-site life detection on Mars and other celestial bodies.

1 of 1

Story summary

Researchers developed a machine-learning technique to identify signs of ancient life in rocks up to 3.3 billion years old.
The study published on November 17 shows the AI system distinguished biotic from abiotic samples with over 90% accuracy.
Its confidence decreases for older samples.
The method detected biogenic molecules in 3.3-billion-year-old rocks from South Africa.
Experts say the technique could aid on-site life detection on Mars and other celestial bodies.