Astronomers are harnessing artificial intelligence to unlock discoveries from decades of space data, with European Space Agency researchers using neural networks to identify hundreds of previously unknown cosmic objects in Hubble's 36-year archive. Pablo Gómez and David O'Ryan deployed Anomaly Match, an AI system trained to detect unusual patterns, revealing galaxies with peculiar shapes that manual inspection missed.

Oxford University's Héloïse Stevance developed a Virtual Research Assistant that slashes supernova signal verification time by 85%, reducing what once took hours of daily human "eyeballing" to minutes. Her team's AI tool processes alerts from five ground-based telescopes scanning for high-energy stellar explosions, freeing researchers for deeper scientific analysis rather than routine data sorting.

Meanwhile, University of Birmingham's Guy Davies employs AI emulators to model stellar evolution, compressing weeks of computation into milliseconds once systems are trained. His colleague Anjali Piette applies similar techniques to exoplanet atmosphere research, connecting spectral data to molecular processes. These advances prepare astronomers for the Vera Rubin Observatory in Chile, which will capture the southern sky in ultra-high definition every 30 seconds for a decade.

The NSF-Simons AI Institute for the Sky (SKAI) in Chicago is developing new tools to handle this data deluge. Unlike generative AI platforms requiring massive resources, these specialized astronomy tools run efficiently on standard hardware—a single GPU suffices for most tasks—making AI-driven discovery both sustainable and scalable.