assign characteristics but how do machines learn and when do they make mistakes. researching these very questions. clearly shows that the horse in the picture has not been recognized for the models entire attention is focused on this text on this copyright tag that really surprised us. we looked at the data set again and saw that many of these horse pictures had a copyright tag on them and that the system doesn't recognize the horses at all which is precisely what you want to avoid or this. so my kids the group machine learning at the berlin for its institute neural networks are fed with huge quantities of data however it's not always clear which parts of the data they use to make their decisions. this is what the researcher wants to find out artificial intelligence isn't as advanced as people think. yes karissa seventy yeah i systems