NASA launched Artifact InSPECtor on September 11, 2026, inviting volunteers to identify errors in space telescope imagery to improve artificial intelligence models.
The project asks members of the public to review spectrographic observations collected by space observatories. Volunteers inspect data gathered by the European Space Agency’s Euclid telescope, built with contributions from NASA, and will review readings from the Nancy Grace Roman Space Telescope starting in early 2027. Both spacecraft use spectrographs to split galactic light into color patterns called spectra, allowing researchers to measure distances, examine stellar compositions, and study supermassive black holes.
Telescope data frequently registers false signals known as artifacts. According to NASA, these unwanted signatures stem from cosmic rays hitting detectors, light bouncing off spacecraft housing, and camera or electronic quirks. Astronomers built AI tools to flag and remove these anomalies, but the algorithms struggle to classify artifacts accurately in readings from newer hardware.
Volunteers categorize these signatures using smartphones, tablets, or computers to correct the training instructions supplied to the machine learning systems. "It’s really cool that we can help teach computers new skills," said nine-year-old Maeve F., a participant in the program.
The refined data will support investigations into dark energy and cosmic expansion. While Euclid gathers light across millions of distant galaxies, the Roman observatory will observe a comparable volume of galaxies across different distances and sky densities after beginning science operations.
