or, even worse perhaps the productivity gains will be high.
This is probably what is going on when we see McD executives say there are basically two economies in the US at the moment, one doing very well for itself and the other experiencing a lot of hardship.
My comment is really a off-topic, its that if AI really does work well it'll put half of us out of work, leaving us to do non-knowledge work that the robots can't do yet.
Your phrasing there is misleading. The article says, "In the first half of this year, business spending on AI added more to GDP growth than all consumer spending combined," and the key is the "added more to GDP growth" part.
Growth in consumer spending was sluggish, while growth in business AI spending was insane, so in terms of how much the economy grew, the rise in AI spending exceeded the rise in consumer spending. Which is pretty amazing, actually. But, as far as the total amount of spending, business AI is at most only a few percentage points of GDP (and that's if you interpret "business AI spending" very broadly), while consumer spending is somewhere around 67-70% of GDP.