Corrective feedback plays a critical role in enhancing writing skills among English as a Foreign Language (EFL) learners, but large class sizes often hinder the provision of personalized feedback. Generative AI tools such as ChatGPT have emerged as promising solutions that offer immediate and individualized feedback for writing. This study examined the effects of ChatGPT-generated feedback in comparison to teacher-generated feedback on Saudi EFL students' writing skills and perception at the university level. A mixed-method approach was employed to ensure comprehensive data collection and analyses. Using the pretest-posttest control group design, 102 participants were assigned to an experimental group, which received ChatGPT-generated feedback, and a control group, which received teacher-generated feedback. Pretests, posttests, and a perception survey were used to collect the data. ANCOVA, with pretest scores as a covariate, and content analysis of perception data were conducted. The results indicated no statistically significant differences in posttest scores between the experimental group and the control group, suggesting that ChatGPT-generated feedback was as effective as teacher-generated feedback. The perception survey revealed an overall positive attitude toward the use of ChatGPT for academic writing. The study provides empirical evidence supporting the effectiveness of generative AI tools in academic writing contexts.