A team at Australia's Commonwealth Scientific and Industrial Research Organisation (CSIRO) used an AI model trained on cat and dog pictures to verify a claim about the location of a black hole.
How the team tested the claim The researchers applied image-recognition techniques to astronomical data tied to a reported breakthrough in black hole observation.
By training the model on everyday images before turning it to the scientific evidence, the team built a check on whether the signal pointed to a black hole in the stated location.
Why the method matters Black hole observations demand strong evidence because researchers infer the objects from indirect signals rather than direct views.
For CSIRO, the project places Australian researchers in the effort to confirm one of astronomy's hardest targets: evidence for a black hole.