Again not much progress on the capabilities front. My head welcomes all the incremental advances in creating consensus on safety and locking down uncontrolled frontier models. My heart yearns for new preprints.
Again not much progress on the capabilities front. My head welcomes all the incremental advances in creating consensus on safety and locking down uncontrolled frontier models. My heart yearns for new preprints.
https://observers.france24.com/en/middle-east/20231018-was-this-photo-of-a-dead-israeli-baby-ai-generated-when-ai-detection-errors-muddle-public-debate
Sure, but those systems are known to be useless. Far as I can tell, Zvi’s comment holds up. Falsely believing that an image was AI generated should not count as “AI involvement”, imo.