Right answer.
This popped up a few months back and I'm copypastaing my own thoughts from there as they still apply"
The Good: AI has the potential to solve an extremely large number of problems that humanity has both now and in the future. Calculations and repetitive tasks, production of resources and oversight thereof that would be much more efficient than a human, solving complex logistical problems where a human could not...and that's just a very few of a multitude of potential uses. It has the potential to make this world so much better and render humans so much freer by giving them the most valuable resource of all - time.
The Bad: However, such utilisation is only as good or bad as the people who utlilise it, and therein lies a big problem - it can also be used to oppress and control in a way that no other civilisation has had at its fingertips until now, due to the sheer amount of elements it can be put in charge of. It can also be used to further inequality and thus civil unrest by putting a lot of creative artists out of jobs while furthering profit for those at the top who control it. The science fiction scenario of it being put in charge of weapons of mass destruction that are then used to destroy us is fanciful (because a decision to remove humans from such decision-making really isn't going to happen because everyone is aware of that exact trope), but it doesn't have to be that direct to cause a great deal of suffering to human civilisation.
So, overall, it's the same as any other tech development - dependent on the humans who control it - only on a much larger scale than tech developments in the past, even more than the steam engine or the Internet. We should embrace it, but we must also be damn careful.