This study evaluates the effectiveness of six major Generative AI (GenAI) text detectors when confronted with machine-generated content that has been manipulated to evade detection. The results show that the detectors' already low accuracy rates (39.5%) drop significantly (17.4%) when faced with manipulated content, with some techniques being more effective at evading detection than others. The limitations of these tools highlight the challenges educators face in maintaining inclusive and fair assessment practices. However, they may support student learning and maintain academic integrity when used non-punitive. The study underscores the need for a combined approach to address GenAI challenges in academia to promote responsible and equitable use of these technologies. It concludes that current AI text detectors have limitations that require a critical approach for any implementation in higher education and highlights possible alternatives to AI assessment strategies. The research explores the reliability of AI text detectors in determining the source of text, the effectiveness of adversarial techniques in disguising AI-generated content, and the comparative performance of AI tools. It also investigates which AI detectors show more promise in terms of accuracy and reliability. The study finds that AI text detectors have significant limitations in detecting AI-generated content, particularly when manipulated using adversarial techniques. These techniques can lead to false accusations, especially against non-native English speakers (NNES), and can disadvantage certain groups in academic settings. The results show that AI detectors have varying levels of accuracy, with some performing better than others. The study also highlights the potential for AI detectors to be bypassed using adversarial techniques, which can lead to undetected AI-generated content and unfair advantages for some students. The findings suggest that the use of AI text detectors in higher education should be approached with caution, as they may not be reliable for determining academic integrity violations. The study concludes that while GenAI tools have the potential to enhance academic practices, their use in assessment must be carefully considered to ensure fairness and inclusivity. The findings suggest that a balanced approach is needed, incorporating both AI detection tools and alternative strategies to promote equitable assessment practices. The study also emphasizes the need for further research to evaluate the effectiveness of AI text detectors and to develop more reliable methods for detecting AI-generated content.This study evaluates the effectiveness of six major Generative AI (GenAI) text detectors when confronted with machine-generated content that has been manipulated to evade detection. The results show that the detectors' already low accuracy rates (39.5%) drop significantly (17.4%) when faced with manipulated content, with some techniques being more effective at evading detection than others. The limitations of these tools highlight the challenges educators face in maintaining inclusive and fair assessment practices. However, they may support student learning and maintain academic integrity when used non-punitive. The study underscores the need for a combined approach to address GenAI challenges in academia to promote responsible and equitable use of these technologies. It concludes that current AI text detectors have limitations that require a critical approach for any implementation in higher education and highlights possible alternatives to AI assessment strategies. The research explores the reliability of AI text detectors in determining the source of text, the effectiveness of adversarial techniques in disguising AI-generated content, and the comparative performance of AI tools. It also investigates which AI detectors show more promise in terms of accuracy and reliability. The study finds that AI text detectors have significant limitations in detecting AI-generated content, particularly when manipulated using adversarial techniques. These techniques can lead to false accusations, especially against non-native English speakers (NNES), and can disadvantage certain groups in academic settings. The results show that AI detectors have varying levels of accuracy, with some performing better than others. The study also highlights the potential for AI detectors to be bypassed using adversarial techniques, which can lead to undetected AI-generated content and unfair advantages for some students. The findings suggest that the use of AI text detectors in higher education should be approached with caution, as they may not be reliable for determining academic integrity violations. The study concludes that while GenAI tools have the potential to enhance academic practices, their use in assessment must be carefully considered to ensure fairness and inclusivity. The findings suggest that a balanced approach is needed, incorporating both AI detection tools and alternative strategies to promote equitable assessment practices. The study also emphasizes the need for further research to evaluate the effectiveness of AI text detectors and to develop more reliable methods for detecting AI-generated content.