AI Model Spots Two Possible Planets Around Distant Star
An artificial-intelligence tool, directed by a non-astronomer, flagged two candidate planets orbiting a star 116 light years away.
Updated on Nov. 20, 2026 in Space

Live Poll
Do you trust scientific findings produced by AI tools used by non-experts?
Pavel Rabtsevich, an artificial-intelligence specialist in Spain with no formal astronomy education, used Claude Code to search NASA TESS data and identify two possible planets orbiting star TIC 4206066. The AI delivered the findings in about two weeks, and TESS astronomers have since added the two signals to their official list of planet candidates.
Why it matters
The case shows how AI tools can help sift large volumes of telescope data for signals that might otherwise go unnoticed. The findings have not yet been peer reviewed, so the planetary candidates still require confirmation by astronomers.
The analysis covered more than 120,000 stars observed by TESS in late 2025 and early 2026, and the star TIC 4206066 showed dimming cycles of roughly 3 and 11 days that matched archival data from 2018 and 2020. Whether these periodic dips reflect two real planets remains unconfirmed pending peer review.
The players
Claude Code
An AI coding tool developed by Anthropic that was used to analyze the telescope data for planetary signals.
Transiting Exoplanet Survey Satellite (TESS)
A NASA space telescope mission that observes stars to detect dips in brightness caused by orbiting planets.
Michelle Kunimoto
An exoplanet researcher at the University of British Columbia in Vancouver, Canada, whose publicly available planet-detection algorithm was used to validate the signals.
The details
Rabtsevich directed Claude Code to sift the TESS data for dips in starlight that could indicate orbiting planets. The AI zeroed in on TIC 4206066, and a planet-detection algorithm written and made publicly available by Michelle Kunimoto, an exoplanet researcher at the University of British Columbia in Vancouver, was used to validate the signals. Rabtsevich published the findings as a preprint on the platform Zenodo on 25 September.
Timeline
2018 and 2020: Archival TESS data recorded the same dimming pattern from TIC 4206066.
Late 2025 and early 2026: TESS observed the more than 120,000 stars later analyzed by Claude Code.
September 2026: Rabtsevich asked the AI model to search for undetected planets.
25 September 2026: Rabtsevich published his findings as a Zenodo preprint.
Big Picture
The discovery extends NASA's TESS mission, whose archived observations are now being mined with AI tools to add entries to its official catalog of planet candidates. It shows how archival space-telescope data can yield new findings when analyzed with methods beyond those of traditional astronomy teams.
For readers following exoplanet research, the case shows that AI tools can now be used by people without astronomy training to flag candidate planets in public telescope archives. Any planetary claim from this work should be treated as provisional until peer review and independent confirmation are completed.
The takeaway
Public telescope archives are open to anyone with the right tools, but candidate planets flagged by AI still need expert confirmation before they are accepted as real. Readers should watch for peer-reviewed results on TIC 4206066 before drawing firm conclusions.
Further reading
Explore more discoveries on the Space section for related coverage of planet hunting and telescope data.
Source note: This article includes information reported by Nature.
Live Poll
Do you trust scientific findings produced by AI tools used by non-experts?



