AI might find another earth before we do!

9 Sept 2026 · 1 min

Watch on YouTube

0:53

In this video, Dan Novaes explains how the massive volumes of data gathered by space telescopes are beyond the capacity of human analysts, and how artificial intelligence is employed to sift through these datasets to detect the minute dimming of stars caused by transiting exoplanets. He describes the transit photometry method, the role of AI in flagging candidate signals, and the subsequent human verification process that confirms true planets while discarding false positives. Novaes emphasizes that as newer telescopes generate even larger data streams, AI’s ability to locate these hidden signals will become increasingly critical, potentially revealing countless undiscovered worlds—including Earth‑like planets—already present in archival observations.

Chapters

Space telescopes collect an unbelievable amount of data, so much that scientists can't manually inspect all of it. That's where AI comes in. When a planet passes in front of a distant star, the star becomes slightly dimmer, sometimes by less than 1%. Hidden inside years of telescope observations, those tiny changes can reveal planets we've never seen before. AI can scan enormous datasets, looking for those patterns, and it has already helped astronomers identify previously missed exoplanets. But AI doesn't simply announce, we found a planet. Scientists still verify the signal, rule out other explanations, and confirm the discovery. AI's job is finding the needle inside an astronomical haystack. And as new telescopes collect even more information, that ability will become increasingly valuable. There could be countless worlds already hiding inside data we've collected. We just haven't noticed them yet. And strangely enough, AI may help us discover another Earth before a human ever sees it.