When starting developing a video game, should one focus on creating the environment (buildings, trees, mountains etc) first or the A.I. (Playable character, NPCs etc)?
The first step is to create a minimum viable product.
A MVP is the absolute minimum which can be considered a playable game. It doesn't need any graphics yet except for some placeholders. It should just accept basic input and implement the most core game mechanics.
Then iterate from there.
The reason is that you need a test-bed for testing out if your game idea actually works. You could waste months developing assets and AI for a game idea which turns out to not be fun at all when actually played and you would be forced to trash it all.
Also, it's next to impossible to design assets and AI "blind" without being able to put them in the game and experience for yourself how they look and feel when actually playing with them. When you have no working prototype game to put these into you might end up wasting countless hours. You will design and polish assets which you will then redesign from scratch when you have finally the chance to see them in the game. And after that you might end up throwing them away anyway because you decide to remove the game mechanic which uses them for not being fun.
Then when you have your prototype, the question "AI first or graphics first" is quite nonsensical because you usually have programmers and artists working on each in parallel. However it is important for them to communicate, because look and behavior of actors in a game need to match. The player will expect that a "bulky brute" enemy behaves differently than a "nimble speedster" enemy. The process can be driven by the programmers ("We are programming an object which behaves such-and-such, come up with how it could look"), by the artists ("We are designing an object which looks such-and-such, come up with a way how it could behave") or it could be a synergetic process where both sit together and brainstorm ideas for new game entities. Whatever process works for your particular team.
You should figure out the story you want to tell, and the purpose of the game. These will guide everything else and give the user something to come back to time and again.
Starting with a story line always helps. Then, the character and the environment are of equal importance. choose the environment next, because that helps determine what the character can do. The character comes last, because the environment has mostly fixed physics