Actor-Critic is a method in reinforcement learning that combines policy-based (actor) and value-based (critic) approaches. The actor decides on actions while the critic estimates the value function to guide the actor's updates. This approach often converges faster and more stably than purely policy- or value-based methods.