Along with the increasing threat of cybercrime, which is predicted to cause losses of up to US$10.5 trillion by 2025 , penetration testing (pentest) has become a crucial strategy for identifying security vulnerabilities. However, the manual pentest process is often time-consuming. This research aims to analyze the role, effectiveness, and challenges of Generative AI (GenAI), specifically GPT-4.1, in accelerating and optimizing the penetration testing process. This research method uses a qualitative approach with a case study on the "PumpkinFestival" VulnHub machine , where GPT-4.1 is integrated into the Kali Linux environment through the ShellGPT tool. The results show that GPT-4.1 can significantly accelerate all stages of the pentest, from reconnaissance to exploitation. GenAI proved effective in analyzing scan results, composing specific payloads, and creating decryption scripts quickly and accurately , while also filling a research gap by evaluating a newer AI model compared to previous studies. The implication is that the integration of GenAI in cybersecurity has great potential to increase the efficiency and effectiveness of security teams in facing increasingly complex threats.