The three claims seem massively overstated and way beyond my understanding of the state of the art. How can AI “break down the positives and negatives” of a pet

The three claims seem massively overstated and way beyond my understanding of the state of the art. How can AI “break down the positives and negatives” of a petition and predict its future effect if implemented or rejected? I’m I guess willing (ableit horrified) to be shown to be wrong, but this looks ridiculous as presented here.