AI Uncovers Earth’s Oldest Life: 3.3-Billion-Year-Old Traces Reveal a New Chapter in Early Evolution

In partnership with

Shoppers are adding to cart for the holidays

Over the next year, Roku predicts that 100% of the streaming audience will see ads. For growth marketers in 2026, CTV will remain an important “safe space” as AI creates widespread disruption in the search and social channels. Plus, easier access to self-serve CTV ad buying tools and targeting options will lead to a surge in locally-targeted streaming campaigns.

Read our guide to find out why growth marketers should make sure CTV is part of their 2026 media mix.

Learn more.

For decades, scientists have searched for decisive evidence of the earliest life on Earth—tiny biological signatures preserved in ancient rocks, chemical traces hidden in minerals, and microscopic fossils shaped by early microbial communities. While some discoveries have hinted at life dating back more than 3.5 billion years, the evidence has often been debated, incomplete, or nearly impossible to verify with traditional scientific tools. But that picture is rapidly changing.

In a groundbreaking new study, researchers have used advanced machine learning techniques to identify the oldest known chemical biosignatures—faint molecular traces of ancient microbes—embedded in 3.3-billion-year-old rocks from South Africa. The findings not only push the timeline of detectable life on Earth further back but also open a new frontier in the search for life beyond our planet.

Published in the Proceedings of the National Academy of Sciences, the research represents a major collaboration led by scientists from the Carnegie Institution for Science, including noted geologist Robert Hazen. By applying machine learning to extremely degraded organic molecules, the team has opened a window into biological processes that occurred billions of years before humans existed.

A Hidden Whisper in Ancient Rocks

The rocks analyzed in this study come from South Africa’s Barberton Greenstone Belt—one of the oldest and best-preserved geological formations on Earth. These rocks hold a rich record of Earth’s early environment, formed at a time when the planet was a molten, hostile world slowly transforming into a cradle for microbial life.

Traditional chemical analysis struggles to detect biological molecules that have been degraded by time, pressure, and heat. Over 3 billion years, most original biomolecules—lipids, fats, sugars, amino acids—vanish, leaving only microscopic fragments behind.

However, using machine learning, the researchers trained algorithms to recognize subtle patterns within these molecular fragments. These patterns act like fingerprints—“whispers,” as Hazen put it—that differentiate molecules formed by living organisms from those created by non-biological processes.

Hazen explained that the machine learning models are capable of detecting patterns human eyes cannot see, enabling scientists to distinguish biological from non-biological molecules with remarkable accuracy, even after billions of years of chemical damage. This represents a revolutionary shift in how scientists identify ancient life.

What the Chemical Signatures Reveal

One of the most astonishing findings concerns evidence of oxygen-producing photosynthesis. Chemical traces discovered in rocks about 2.5 billion years old suggest that early marine bacteria were using sunlight to produce energy—and oxygen—much earlier than previously believed.

Until now, the Great Oxygenation Event was thought to have begun around 2.4 billion years ago. The new findings indicate that the roots of oxygenic photosynthesis may extend further back in time. These early photosynthetic microbes would have played a key role in reshaping Earth’s atmosphere, paving the way for the evolution of more complex, oxygen-dependent life forms.

This discovery adds a new dimension to the history of life on Earth. Physical fossils from the early period—such as 3.5-billion-year-old stromatolites from Australia and South Africa—offer valuable insights, but such fossils are exceedingly rare. Chemical evidence, especially when enhanced by AI, provides a far more precise way to understand early microbial ecosystems.

The new method reveals not just whether life existed but also what kinds of metabolic processes ancient microbes used. This allows scientists to reconstruct ancient environments, climate patterns, and the biological innovations that shaped early Earth.

Why This Breakthrough Matters for Astrobiology

Beyond rewriting Earth’s history, this research carries enormous implications for the search for life beyond our planet.

According to co-author Anirudh Prabhu, their approach effectively doubles the age of reliable biosignatures, meaning scientists can now potentially identify traces of life in rocks that are more than 3 billion years old—even if the molecules have been heavily altered or damaged.

NASA, recognizing its potential, has begun funding further development of the technique.

Robotic missions on Mars, including the Perseverance rover, already collect rock samples that will someday be returned to Earth. Machine learning–based biosignature detection may be one of the most powerful tools for analyzing those samples. If life ever existed on Mars—even as simple microbes—its chemical traces may still be present in ancient rocks.

Beyond Mars, the technique could also be applied to samples from icy worlds such as:

• Enceladus, which sprays organic-rich water vapor from its subsurface ocean

• Titan, which contains complex organic chemistry

• Europa, another prime candidate for harboring a subsurface ocean

If these bodies ever hosted microbial life, the molecular traces left behind may be fragmentary—but with this new method, they may finally be detectable.

Hazen noted that the team hopes to use this approach for sample-return missions in the near future. Their long-term vision is to build a universal framework capable of distinguishing biological molecules from abiotic ones anywhere in the solar system.

A New Era of Detecting Life in Deep Time

This study marks a major turning point in both geology and astrobiology. For the first time, AI is being used to interpret chemical signatures older than any physical fossil, revealing biological activity that cannot be seen by conventional tools.

By extracting patterns from molecular “noise,” machine learning allows scientists to reconstruct the chemical behavior of ancient microbial communities, identify early metabolic pathways, and refine our understanding of life’s origins.