Artificial intelligence (AI) is rapidly integrating into various sectors, including education. While AI offers numerous potential benefits in education, such as personalized learning and automated administrative tasks, concerns regarding its ethical implications are emerging. This study focuses on three key areas of concern: security and privacy issues related to data collection and usage; the potential for AI to diminish human decision-making capabilities; and the impact of AI on human laziness and reduced engagement. The widespread adoption of AI in education, fueled by significant investments (projected to reach USD 253.82 million from 2021 to 2025), necessitates a thorough understanding of its potential downsides. Many researchers highlight the positive aspects of AI in education, but often overlook the ethical dilemmas and unintended consequences. This study addresses this gap by exploring the negative impacts of AI on university students in Pakistan and China, aiming to inform the responsible development and deployment of AI in educational settings.

The literature review examines existing research on the ethical considerations of AI in education. It explores three levels of ethical challenges: challenges inherent in the technology's design and development, including issues of bias, data misuse, and lack of transparency; challenges faced by educators in integrating and utilizing AI effectively in the classroom, focusing on concerns about security, implementation, and the potential for AI to replace rather than augment human capabilities; and challenges faced by students, including concerns about privacy, data security, and the potential for AI to reduce their critical thinking skills and foster over-reliance on technology. The review also examines existing ethical guidelines and principles for developing AI systems in education, such as those proposed by Aiken and Epstein (2000), highlighting the need for a comprehensive and robust ethical framework to address the complexities of AI integration. The review highlights the lack of a universally agreed-upon framework for addressing the ethical issues surrounding AI in education, emphasizing the need for further research and the development of effective regulatory mechanisms.

This study employs a quantitative research design grounded in a positivist philosophy. Data was collected using a purposive sampling technique, surveying 285 university students in Pakistan and China. A questionnaire was used to gather data on students' perceptions and experiences with AI. The Partial Least Squares Structural Equation Modeling (PLS-SEM) technique, specifically using SmartPLS software, was employed for data analysis. The methodology addresses potential common method bias (CMB) by checking Variance Inflation Factor (VIF) values. The reliability and validity of the data were assessed using Cronbach's alpha, composite reliability, Average Variance Extracted (AVE), Fornell-Larcker criterion, Heterotrait-Monotrait ratio of correlations (HTMT), and cross-loadings. The demographic profile of the respondents, including gender, nationality, age group, and program of study, is analyzed to understand the sample characteristics. A structural model is developed to examine the relationships between AI use and the three key dependent variables: loss in human decision-making, human laziness, and security and privacy issues.

The key findings of the study demonstrate significant relationships between AI use and the three dependent variables. Specifically: 1. **Artificial Intelligence and Loss of Human Decision-Making:** The study found a statistically significant positive relationship (β = 0.277, p < 0.05) between AI use and the loss of human decision-making capabilities among students. This indicates that increased reliance on AI systems for decision-making is associated with a decline in students' ability to independently analyze situations and make informed choices. 2. **Artificial Intelligence and Human Laziness:** A strong, statistically significant positive relationship (β = 0.689, p < 0.05) was found between AI use and increased laziness. This suggests that the automation of tasks by AI systems may lead to reduced effort and engagement on the part of students. This finding highlights a significant negative impact of AI on student learning and personal development. 3. **Artificial Intelligence and Security and Privacy Issues:** A significant positive relationship (β = 0.686, p < 0.05) was observed between AI use and concerns about security and privacy. This underscores the vulnerability of students' data to misuse and the importance of addressing security and privacy issues associated with AI implementation in educational settings. The importance-performance matrix analysis (IPMA) revealed that the importance of addressing security and privacy concerns was highest (74.6%), highlighting the critical need for improved security measures. Multi-group analysis showed no significant moderating effect of gender or nationality on the relationships between AI use and the three dependent variables. The model's goodness-of-fit was confirmed by a SRMR value of 0.006, indicating a good fit. The predictive relevance was assessed using Q² values, indicating a good predictive power for the model in explaining the dependent variables.

The findings of this study highlight the significant challenges associated with the integration of AI in education. The positive relationships found between AI use and the loss of decision-making skills, increased laziness, and security and privacy concerns underscore the need for a cautious and responsible approach to AI implementation. The study's findings are consistent with previous research highlighting the ethical implications of AI in various contexts. The high importance rating for security and privacy issues in the IPMA analysis reinforces the urgency of developing robust security protocols and addressing potential privacy violations associated with AI systems in education. The lack of significant moderating effects of gender and nationality suggests that the observed trends are consistent across different demographics. The study's findings contribute to a growing body of literature calling for greater attention to the ethical implications of AI and the need for a balanced approach that leverages the benefits of AI while mitigating its potential risks.

This study demonstrates that while AI offers potential benefits for education, its implementation must be approached cautiously. The significant negative impacts on decision-making, increased laziness, and security/privacy concerns necessitate preventive measures and ethical considerations. Future research should explore mitigating strategies and the development of comprehensive ethical guidelines for AI in education. Further investigation into diverse cultural contexts and educational settings would also enhance our understanding of this complex issue.

The study's primary limitation lies in its focus on three specific ethical concerns. Other ethical issues related to AI in education were not investigated. The generalizability of the findings may be limited by the sample's geographic location and the use of a specific methodology. Future studies could incorporate a broader range of ethical concerns and employ mixed-methods approaches to gain a more comprehensive understanding of the issue.