Post #1876785

Expect big debates over the use of "open source" datasets in AI development. On one hand, LAION published thousands of examples of illegal CSAM and got them incorporated into AI training models- which is a huge fail for them and everyone else who trained on it. On the other hand, the Stanford Internet Observatory was able to detect and alert people to this because it was an open dataset. https://www.404media.co/laion-datasets-removed-stanford-csam-child-abuse/