Editorial illustration for Anthropic’s Claude Science crafts first complete ultraviolet sky map
Claude AI Creates First Complete UV Sky Map
Anthropic’s Claude Science crafts first complete ultraviolet sky map
Anthropic has put its Claude models to work on a problem astronomers have been stuck with for decades: a complete picture of the sky in ultraviolet light. The company's Claude Science project just finished stitching together the first full UV map of the entire sky, pulling in data from multiple space missions since UV light can't be measured from the ground. The ozone layer blocks it before it reaches a telescope on Earth, which is why NASA's GALEX mission, the closest thing to a precedent, only managed to cover about two-thirds of the sky and left out bright star-forming regions entirely.
Johns Hopkins astrophysicist Brice Ménard, who worked with Anthropic on the project, laid out the process on the company's website. UV light matters because it picks up dust lit by nearby stars, the kind of faint structure that shows up around stellar nurseries or the leftover rings of supernovae. Building a usable map meant downloading archives from several missions, calibrating them against each other, and filling in the gaps where no satellite had looked. That last part is where Claude's AI agents did the heavier lifting, and it's also where the results get interesting.
Anthropic's Claude Science just mapped the entire sky in ultraviolet light for the first time. The project is supposed to show how AI can take over tedious data work that researchers would otherwise never get around to.
Why this matters
This is a useful test case for what AI actually earns its keep doing right now: not generating new physics, but finishing a job humans started and abandoned. GALEX left a third of the sky unmapped because nobody wanted to grind through the archival data by hand. Claude Science did, and Ménard's willingness to put his name on the result at Johns Hopkins matters more than the Anthropic press framing around it.
For researchers, the lesson is narrower than the headline suggests: this works when there's a known physical signal (UV-lit dust, stellar remnants), a bounded dataset, and a domain expert checking the output. Founders pitching "AI does science" should note that's exactly the kind of scaffolding missing from most of those pitches. We'd watch for whether this map gets used by other astronomers, cited in papers, built on.
That's the real test. A press release calling something "the first complete UV sky map" is a claim; a few years of other people using it without having to redo it is evidence.
Related Reading
Browse all Anthropic news →
LLMs & Generative AI ChatGPT's 'Nerdy' tweak rewards goblin metaphors in answers, study finds
LLMs & Generative AI Google to End Free Access to Gemini AI for Non-Subscribers
LLMs & Generative AI AI Engineers Face Rising Costs, Need New Strategies for Efficiency
Related Topic Claude gains shared context in Excel, PowerPoint; Microsoft adds Copilot Cowork
AI Tools & Apps
Related Topic Trump cracks down on Anthropic after Amazon tip; staff largely foreign
Open Source
From Archives Anthropic’s Claude Creates Animated Videos and Live Data Dashboards From Text
From Archives Fired OpenAI Researchers Warn of Lost Safety Tool for AI Reasoning
From Archives Anthropic Suspends Accounts for Mistreating Claude Under New Terms
Policy & Regulation
From Archives Anthropic's Claude Haiku 5.5 Cuts Input Token Cost to USD 0.10 Per Million
Open Source
Common Questions Answered
Why couldn't astronomers create a complete ultraviolet sky map from Earth-based telescopes?
Earth's ozone layer blocks ultraviolet light before it can reach ground-based telescopes, making it impossible to measure UV light from the surface. This is why space missions are required to capture ultraviolet data from the entire sky.
What did Anthropic's Claude Science project accomplish with the ultraviolet sky map?
Claude Science stitched together the first complete ultraviolet map of the entire sky by integrating data from multiple space missions. This achievement filled a significant gap left by NASA's GALEX mission, which had only mapped two-thirds of the sky because researchers found manually processing the archival data too tedious.
How does the Claude Science UV mapping project demonstrate AI's current practical value?
The project shows that AI excels at completing tedious data work that human researchers would otherwise abandon or never prioritize. Rather than generating new scientific discoveries, Claude Science finished a job that astronomers had started but left incomplete due to the labor-intensive nature of processing archival data.
Why was NASA's GALEX mission unable to complete a full ultraviolet sky map?
GALEX only mapped approximately two-thirds of the sky because manually grinding through the remaining archival data was too time-consuming and labor-intensive for researchers to undertake. This left a significant portion of the ultraviolet sky unmapped until Anthropic's Claude Science project completed the task.
