In a major win for both alien hunters and artificial intelligence stans, researchers at the University of Bern have developed an AI that’s identifying where the next Earth-like planets might be hiding. And according to their new study published in Astronomy & Astrophysics, the algorithm has already flagged 44 star systems that could host planets with conditions favorable to life.
No, the AI hasn’t confirmed these planets exist—but it’s narrowed down the cosmic haystack astronomers have been poking through for decades. That alone could be a turning point in the search for life in the universe.
“This is a significant step,” said Dr. Yann Alibert, one of the researchers behind the project. “It enables predictive studies like ours that simply weren’t possible before.”
See also: Is There Life Beyond Earth? Astronomers Detect Potential Sign of Life on Distant Planet
Wait, Did an Algorithm Just Outsmart Our Telescopes?
Kinda, yeah. Spotting exoplanets—planets outside our solar system—is notoriously difficult. They’re tiny, cold, and don’t produce their own light, making them almost invisible next to the stars they orbit.
So far, scientists have only confirmed about 5,800 exoplanets, and most of them are just blurry data points. That makes it hard to train any AI model, which usually requires massive datasets to learn patterns.
To work around that, the Swiss team fed their algorithm a simulated cosmos instead. Using something called the Bern Model of Planet Formation and Evolution, they created detailed synthetic star systems based on how planets form in real life—from protoplanetary dust discs all the way to full-blown planetary systems.
The AI then learned to identify subtle patterns in planetary formation, especially focusing on the innermost detectable planet in each system—its mass and orbital period ended up being the biggest clues.
See also: No, Scientists Didn’t Resurrect the Dire Wolf — But What They Did Do Is Even Weirder
44 Systems, One Big Clue in the Search for Another Earth
Armed with its new knowledge, the AI scanned 1,600 real star systems containing at least one known planet and one of three common star types: G-type (like our Sun), K-type, and M-type. It predicted that nearly four dozen of these are likely hiding an Earth-like planet we haven’t seen yet.
That’s huge. Because now, astronomers know exactly where to look next.
See also: (VIDEO): Scientists Capture the First-Ever Footage of a Live Colossal Squid in the Wild
It’s Not a Crystal Ball—Yet
The AI isn’t flawless. It failed to account for some observed features, like the correlation between “Super Earths” and “Cold Jupiters” in Sun-like systems. And the fake planets it trained on tended to orbit closer to their stars than real ones do.
Still, the researchers aren’t trying to replace telescopes—just give them a better aim. The tool doesn’t need to be perfect. It just needs to be directional.
After all, the cosmos is big. Like, mind-meltingly big. Even the most advanced telescopes are still just guessing where to look. If AI can take that guesswork down even a few notches, the odds of finding another Earth—or even alien life—go way up.
See also: NASA Astronauts’ Stunning Transformation After 9 Months in Space – Before and After Photos
So… Are We Alone or What?
That’s still the big question. But with AI accelerating the search, we might not be wondering forever.
Whether or not these 44 systems turn up habitable planets remains to be seen. But one thing is certain: the hunt for life in the universe just got a powerful new sidekick. And it’s not human.
