The MiniMax algorithm is a recursive algorithm which the AI assesses all the possible outcomes and assigns a score based on the terminal state. The role of the maximizer is to score the highest positive amount of points as the minimizer trying to score the most negative amount of points. This maximizer-minimizer role exists solely in the algorithm and not to the game as a whole.

In this algorithm, the computer plays the role of the maximizer and the human is the minimizer. When it is the computer's turn, the algorithm will find all the empty squares and loop through all the spaces. The computer will place a marker in the first empty space, then alternate with the player's marker to the next empty space, then alternates back to the computer, and so on. This meta process occurs until it reaches a terminal state.

Upon reaching the terminal state and it is a win for the computer, the score will return a positive 10. If it's loss for the computer, it will return a negative 10. If that iteration does not have a terminal state, the algorithm will continue running deeper until it reaches that state. The score then passes up to the original first square that the computer marked. Subsequently, each square now has a score that the computer can evaluate to calculate the best move. In a sense, the computer can calculate and see all the possibilities of the end of the game and can pick the best path to win. It evaluates all these possibilities on every turn and follow destiny to either a win or a minimum, a draw state. This is the beauty of the MiniMax Algorithm.

There are a ton of youtube videos, articles, and repositories of some incredible renditions of this projects so stop by your local search line on Google to find more

Some recommended resources are: