Deepfake dangers, AI image opt out, and controlling your urges
Smashing Security
00:00
Do You Care if You're Not Recognized?
Some of the worst images that have ever been posted on line are also included in the data set. Such stuff that was very clearly illegal or grossly offensive is being used to train these esaes. There's no accountability for the content as a result, you know, there is a moral imperativeness. Laon siht doesn't even go into detail about thet the not safer work and violent images that appear in the dataSet. It does say that, er, it does not contain images that may be disturbing to viewers, which is untrue. But links in the data Set can lead to images that are disturbing, ordiscomforting, dependent on the filter or search method employed
Transcript
Play full episode
