The questions you should ask are:

1) Does your opponent also have AI?

2) Does your opponent have the same level of AI as you do? If not, then is it superior or inferior?

3) Does your opponent have nuclear weapons? Is it mutually assured destruction? If not, then are you at a disadvantage?

Of course, if your opponent does NOT have AI and has NO nuclear weapon a war is definitely more acceptable to you. If your opponent has superior AI and more nuclear weapons, then a war is less acceptable to you.

The more interesting scenario would be a war between two opponents who are equally matched. Then, it boils down to who can produce more war robots. In other other words, who can outlast the other. As soon as this comes into the calculation, then you have to destroy the industrial base for war robot manufacturing. Whether you use missiles or bombs, there would be human casualties. Then it is no different than any current war. Things will quickly escalate and if both parties have nuclear weapons it would eventually lead to a nuclear exchange.

The old scenario of Soviet Union's armoured divisions invading Western Europe...The West not able to withstand a Soviet assault...leads to the West hiring tactical nukes to halt the Soviet advance...leads to Soviet retaliatory tactical nukes...leads to a full nuclear exchange. The same results and consequences but with smarter weapons.

AI would only make a HUGE difference if it is a war between the likes of the US and Iran (or the US and ISIS or the US and the Taliban). 

If the US and Russia get into such a war and Russia is falling way behind in AI...but with enough nuclear weapons, the outcome would still be the same.

An even more interesting scenario would be a war between China and the US and China is way ahead in AI. As the US starts to lose such a war, the US would definitely be tempted to use nuclear weapons. With only a fraction of the nuclear weapons the US possesses, how should China react?