Filling the governance vacuum related to the use of informat...

Filling the governance vacuum related to the use of information commons for AI training

A report investigating the governance challenges of using openly licensed content for AI training, with findings and recommendations for improving AI dataset governance.

openfuture.eu

RelatedInsightsHighlights

Datasets as Imagination

Lila Shroff joinreboot.org

Models All The Way Down

knowingmachines.org

Differing interpretations on the potential of AI art aside, this thread is well reasoned and covers 3 key points we've been trying to make clear for a long time 1) Training data is unlikely to be considered a copyright infringement due to the specifics of how training works. Memorization is very rare and considered a... See more

Mat Dryhurst x.com