Spatial target speaker extraction isolates a desired speaker’s voice from a mixture of multiple speakers using spatial information, such as the direction of arrival (DoA). Although recent deep neural network-based discriminative methods have shown significant performance improvements, the potential of generative approaches, such as generative adversarial networks (GAN), remains largely unexplored for this problem. To the best of our knowledge, we propose the first GAN-based method that leverages spatial information to generate the target speaker’s speech. Additionally, by conditioning the GAN on intermediate features from a discriminative spatial filtering model, along with the DoA, we enable steerable target speaker extraction with a spatial resolution of 5◦, outperforming state-of-the-art discriminative methods in both objective and subjective metrics.