← All research storiesSpace

Peer-reviewedSpace

An AI searched billions of sky objects for “cosmic magnifying glasses,” and telescopes confirmed 70 of its picks

A neural network flagged likely gravitational lenses in the biggest 2D map of the universe; follow-up with a giant telescope confirmed 70 of 76 candidates.

How far alongPeer-reviewedReviewed by independent experts for a journal before publication. Peer review is a quality check, not a grade of evidence strength: confidence still depends on study design, limitations, replication and the wider evidence.
How soon could it reach you?Long-term scienceBasic research that builds knowledge. Real-world uses, if any, are far off.
Where it comes fromThe Astrophysical Journal Supplement SeriesLawrence Berkeley National Laboratory; NSF NOIRLab; Santa Clara University; UC Berkeley · Oct 7, 2026Read the original →

What the scientists did

When a massive galaxy sits almost exactly in front of a more distant one, its gravity bends the light from the far galaxy, stretching it into arcs or rings. These rare lineups are called gravitational lenses, and they are hard to spot by eye among billions of objects.

Astronomers trained a type of AI called a residual neural network to search the DESI Legacy Imaging Surveys, a map of nearly 4 billion objects made from 5.6 trillion pixels. The AI picked out likely lenses. The team then pointed the MUSE instrument on the Very Large Telescope in Chile at 76 candidates to measure how far away the front and back galaxies really are.

Words to know

Gravitational lens

A massive object, like a galaxy, whose gravity bends and magnifies light from something farther away behind it.

  • A glowing ring around a galaxy (an “Einstein ring”)
  • Arcs of light stretched around a galaxy cluster
  • Several copies of the same distant galaxy

Neural network

A computer program that learns to recognize patterns by studying many examples.

  • Spotting faces in photos
  • Picking out arc-shaped lens features in telescope images

Redshift

How much an object’s light has been stretched toward red by the expanding universe, which tells astronomers how far away it is.

  • Nearby galaxies have small redshifts
  • Very distant galaxies have large redshifts

What they found

Of the 76 AI-picked candidates, 70 turned out to be real lenses and only 6 were not. For 55 systems the team measured the distance to both the front galaxy and the lensed galaxy behind it; for 15 more they measured the front galaxy’s distance. Some systems were unusually complex, with more than one background galaxy being lensed.

70 of 76AI-picked candidates confirmed as real lenses
~4 billionobjects in the sky map the AI searched
5.6 trillionpixels in the map

Why it matters to you

This won’t change your daily life, but it shows how AI can act as a tireless first-pass searcher, sifting billions of images so human scientists can focus their limited telescope time on the best targets. That same teamwork between AI and people is used in medicine, farming and many other fields.

Where this could lead

Better maps of dark matter

Confirmed lenses let scientists weigh galaxies, including the invisible dark matter inside them.

Ongoing basic research

Smarter sky searches

Similar AI tools will be needed to sort the flood of images from new telescopes like the Vera C. Rubin Observatory.

Already starting

These are possibilities the research points toward, not promises. Most early findings take years of testing before they reach everyday life, and some never do.

Why scientists care

Gravitational lenses act like natural zoom lenses that let astronomers see faint, distant galaxies and measure dark matter and the expansion of the universe. A large, carefully confirmed set of lenses is a resource researchers can use for years.

Keep in mind

The AI only suggests candidates; each one still has to be confirmed with expensive telescope measurements. This study checked 76 of the roughly 3,500 candidates found earlier, focusing on the southern sky, so most candidates are still unconfirmed.

How far along is the science?

  • Peer-reviewedReviewed by independent experts for a journal before publication. Peer review is a quality check, not a grade of evidence strength: confidence still depends on study design, limitations, replication and the wider evidence.
  • PreprintA manuscript shared publicly, often before or alongside journal review. Check its separate journal and review status (each story says what we know). Results may change.
  • Conference talkPresented to other scientists at a meeting. Usually short and early; a full paper may come later.
  • Agency reportPublished by a science agency such as NASA, NOAA or NIH. Reviewed internally, but not by an outside journal.
  • Press releaseAn announcement from a university or company. Useful, but always check the research it describes.

How soon could it reach you?

  • Could reach you soonCould show up in products, services or advice soon. Our rough editorial guess is about 1 to 3 years, not a forecast from the research.
  • A few years outPromising, but needs more testing or engineering first. Our rough editorial guess is several years, with a lot of uncertainty.
  • Long-term scienceBasic research that builds knowledge. Real-world uses, if any, are far off.
  • Small stepA small update to earlier work. Useful to researchers, but not a change you would notice yet.

Every story links to its original source: the paper itself, or the conference program when no paper is out yet. Explanations are written with AI, usually from the paper’s abstract (each story says what it was written from), then checked against the source. We only use journals, recognized preprint servers, science agencies, scientific conferences and established science news outlets.