1. Introduction

In the 21st century, artificial intelligence (AI) technology has developed rapidly and has been applied across a wide range of industries within just a few decades. Unsurprisingly, AI has also had a profound impact on the education sector worldwide. In particular, the changes driven by the development of AI technologies have recently become a central concern for universities and colleges. On the one hand, AI offers numerous advantages for advancing higher education. For professors and universities, AI can assist with instructional design, student support, and administrative tasks. For students, AI is increasingly integrated into their learning process, often serving as a resource provider or aiding in assignment completion.

On the other hand, the integration of AI does not solely lead to positive outcomes. Its application in higher education also presents significant challenges. For instance, AI-driven instructional design may lack personalization, potentially neglecting the diverse needs of students. At the same time, students' reliance on AI tools to complete academic tasks raises important concerns regarding fairness, academic plagiarism, and the overall quality of the learning experience. In response, many colleges and universities have to respond to these challenges, and have begun to formulate policies regulating the use of AI. In recent years, these universities have consciously banned the use of AI tools or imposed restrictions to mitigate associated risks.

This research aims to examine how universities in various regions are developing policies to address these challenges and how they can implement AI in higher education in a responsible and constructive manner. Eventually, this study contributes to the ongoing discourse on the on evolving role of universities and the higher educational landscape and explores how they can catch the opportunities that AI brings about and adapt to remain innovative, inclusive, and relevant.

2. Opportunities and potentials

With the advancement of technology, the innovation of Artificial Intelligence has brought numerous opportunities for the development of higher education. Whether in quality teaching or university administration, or in the learning process of students, the application of AI tools has demonstrated various potential in this area of development.

2.1. AI-driven feedback and assessment in higher education

The early application of AI in education was primarily reflected in providing immediate feedback to students in higher education, which could potentially improve the efficiency of teaching. As Kelum et.al highlighted, feedback must emphasize fairness, transparency, and realism [1]. Such assessment and feedback can, to a certain extent, enhance students' academic performance and foster motivation for students' learning. Artificial Intelligence has the potential to enhance both the efficiency and practicality of implementing assessment methodologies within educational evaluation processes. The AI-driven StudiAse System and the advancement of intelligent software for online exam selection serve as illustrative examples of how Artificial Intelligence is being employed to enhance student assessment and feedback mechanisms [2,3]. AI-driven systems can automatically grade students' work and exams, thereby reducing teachers' workload. As a result, real-time feedback can be provided, improving learning efficiency and personalization. As Miguel noted, Automated Performance Enhancement (APE) can assist professors in grading papers, and Intelligent Tutoring Systems (ITS) can offer assessment and feedback while creating personalized and adaptive instruction for students. Moreover, AI can automatically type and quantitatively grade, allowing for both grading and self-assessment of students' assignments. Such a grading system can drive personalized guidance and immediate feedback, thereby enhancing the teaching quality in colleges and universities.

2.2. Enhancing learning with AI

Artificial Intelligence holds significant potential in supporting the learning processes of university students, particularly through its role as an auxiliary educational tool. The ability of AI to access information enables students to use AI to ask questions [1]. OpenAI can provide answers quickly, which is believed to enhance students' desire and motivation for learning. Students can use the answers provided by ChatGPT as a starting point and further enhance the answers through their knowledge and ideas. It can be seen from this that AI tools, such as ChatGPT, can enable students to quickly generate output and reach the expected level in the process of their higher education learning

AI can also demonstrate its capabilities in the area of reinforcement learning in the context of changes in the school. Artificial Intelligence presents higher education institutions with the potential to reimagine and enhance the flexibility and accessibility of the learning environment. Technological advancements have created opportunities for both students and teachers. For instance, the emergence of both online and traditional learning platforms has enabled a more flexible digital learning environment [4]. Artificial Intelligence supports the digital dissemination of educational content and the exchange of information, enabling efficient and secure communication between students and instructors, while simultaneously overcoming barriers related to distance and language. These technologies have supported the development of various blended learning models. Students are able to customize their own learning pace and set individual goals, while universities can offer courses more flexibly and provide a greater variety of course options. To adapt to digital education and enhance competitiveness, universities must shift from traditional methods to more digitally integrated approaches. For instance, students now benefit from extensive access to online educational resources, including digital lectures, virtual classrooms, collaborative teaching initiatives, simulated practice environments, and educational applications. Moreover, the integration of emerging technologies--such as augmented reality (AR) and virtual reality (VR)--into the higher education sector has the potential to substantially enhance both the quality and progression of educational delivery and learning experiences.

2.3. Big data in university management

The emergence of Big Data in the development of AI technology enables the handling of large volumes of data with varying structure [5]. Big Data possesses powerful data analysis and processing capabilities, which provide significant potential and opportunities for the management and operation of universities and colleges. In higher education, Big Data encompasses a vast array of both vertical and horizontal data related to students. This includes detailed information on teaching and learning activities, such as social connections, intentions, and goals within the learning process. All of this data can be tracked, allowing institutions to monitor student performance and analyze course evaluations. By analyzing student performance and skill levels, Big Data tools can facilitate the creation of personalized learning experiences. For instance, predictive analytics enabled by Big Data can help institutions enhance student performance and educational quality by collecting and interpreting data throughout the learning process to meet individual student needs. The University of Otago's enhanced Analytics Framework illustrates how universities can employ Big Data through descriptive, predictive, and prescriptive analytics to anticipate potential challenges. This highlights a valuable opportunity to improve university management by timely adjusting the administrative policies and formulating responsive plans.

3. Challenges for universities in responding to Artificial Intelligence

As previously mentioned, the growth of AI technologies has created new avenues for innovation and improvement within higher education. However, alongside the progress of AI technology, improper usage, challenges faced by institutions in implementing AI software, and vulnerabilities within AI systems may also lead to negative impacts and present significant challenges for higher education. These influences could prevent the schools from handling the potential for development brought by the AI.

3.1. Academic integrity in the age of AI

The emergence of Artificial Intelligence has progressively transformed traditional academic practices into hybrid or fully digital formats. While this shift holds considerable potential for advancing higher education, it has also given rise to growing concerns regarding academic misconduct. For instance, the challenge to academic integrity becomes more pronounced when tools like ChatGPT are used in academic work such as essay writing. Students may also use ChatGPT to refine and improve answers multiple times before submitting assignments, thereby creating polished outputs that obscure original authorship [1,6]. These practices, AI-facilitated academic practices, present serious threats to academic integrity and have emerged as a central issue in current evaluation processes across higher education institutions.

3.2. AI inaccuracies in academia

The ability of AI tools like ChatGPT to respond quickly to users' questions and demands often results in the provision of inaccurate information. Such inaccuracies can mislead both professors and students in higher education, potentially leading to serious consequences. A typical example is ChatGPT generating incorrect or fabricated citations during literature searches, many of which cannot be found or do not exist on academic platforms. As a result, both students and faculty are at risk of using false or misleading information, whether they are writing papers or conducting teaching, or research. This presents a significant challenge for universities in maintaining academic quality. Moreover, if students graduate having relied on correct and illogical information, and this leads to failures in their professional work, it could severely undermine public trust in academic qualifications, scholars, and universities.

3.3. AI and the quality of academic learning

In higher education, the knowledge imparted by universities is typically specialized within specific fields, such as economics and philosophy. However, AI tools often lack such domain-specific expertise and instead focus more on integrating general or widely accessible knowledge. In higher education, there is often greater emphasis and importance placed on in-depth insights into professional knowledge in specific fields, critical analysis, and guidance. AI cannot fully meet students' expectations and demands for high-quality professional knowledge, which often poses challenges to the students' learning process [7].

Moreover, higher education emphasizes not only professional knowledge but also the development of effective communication and collaboration skills. As AI language models, current AI tools lack genuine interpersonal communication capabilities and cannot fully replicate the nuances of human interaction. As a result, their ability to support group discussions and other collaborative activities is restricted, posing a challenge to the comprehensive educational experience that universities aim to provide.

In terms of learning behaviors, a more pressing challenge for colleges and universities is growing student dependency on AI. AI tools can rapidly generate tangible results, such as written assignments or explanatory texts, which may discourage students from engaging in deeper cognitive efforts. For instance, students may avoid conducting their analytical processes, increasingly relying on AI-generated content. Such dependency may erode students' motivation for self-directed learning and impede the cultivation of key academic competencies, including problem-solving, critical thinking, creativity, reasoning, and reflexivity [8].

3.4. AI-driven inequality

The use of AI tools, such as ChatGPT, has the potential to exacerbate inequality and unfairness. One notable area where this emerges is in academic assessment. ChatGPT can create written assignments that are often indistinguishable from those produced by students. In some cases, university professors find it difficult to determine whether an assignment has been completed by the students or generated using AI tools. As a result, students may achieve higher grades without genuine effort or critical thinking, thereby outperforming their peers who have completed their work independently. This creates biases in teaching evaluations and undermines the fairness of academic assessment.

In addition, the emergence of AI may exacerbate educational inequality due to the digital divide. Drawing on Pierre Bourdieu's theory of capital, it becomes evident that individuals in regions with relatively limited economic and social resources are less able to access and utilize the Internet and AI technologies as effectively as those in more affluent areas [9]. Consequently, students in under-resourced communities may be unable to fully benefit from the educational opportunities enabled by AI technologies, thereby further deepening existing disparities in educational outcomes.

4. Advice and implementation of AI technology

In response to the complex opportunities and threats posed by AI within the realm of higher education, universities and colleges need to take proactive measures to address the emerging issues while also exploring appropriate and effective ways to utilize AI tools in order to harness their full potential for advancing educational development.

4.1. Designing AI-resistant assessments

The application of AI can affect students' motivation to learn and limit their ability to develop their critical thinking skills. In response to these challenges, universities can design alternative assignment formats and evaluation strategies that stimulate students' enthusiasm for learning while minimizing their reliance on the use of AI. For instance, written assessments, essay writing, and short-answer questions are particularly susceptible to AI-generated responses. In contrast, formats such as multiple-choice questions (MCQs), matching exercises, and performance-based records can be regarded as being able to resist the influence of AI tools. An effective strategy to address this issue is to adopt diverse forms of assessment, such as reverting to traditional closed-book, paper-based examinations or developing personalized, targeted assessment methods generated by AI. The schools could also integrate traditional approaches with AI-based tools to create a more diversified and adaptive assessment framework.

Additionally, institutions can increase opportunities for collaborative learning, including group discussions and presentations. These types of assignments can, to some extent, reduce students' dependence on using AI tools to promote critical thinking and independent learning. Furthermore, incorporating oral or on-site demonstrations and asking students to analyze materials without relying on lengthy texts or visual prompts can also support the development of authentic learning outcomes [10].

4.2. Guidelines for ethical AI use

Employing AI technologies in academic assessments may threaten the integrity of the evaluation process and challenge the fairness of academic outcomes, thereby necessitating the establishment of clear academic policies and procedures. Strict and comprehensive academic integrity policies and procedures are essential to ensure the ethical use of AI tools for academic purposes. For example, in cases of plagiarism, if students reproduce AI-generated content without appropriate attribution, such actions constitute academic misconduct [11].

Secondly, universities should provide training programs for both students and teachers. These programs can help enhance teachers' leadership. Teachers' leadership is of great significance to students' behavior, and it influences their moral development. Educators should demonstrate compassion, motivation, and enthusiasm to guide students in developing critical thinking and ethical awareness, while also encouraging students to use AI tools responsibly and morally in academic works.

In addition, universities also need to continuously monitor and evaluate the impact of AI on higher education. The effectiveness of academic integrity policies can be assessed through feedback collected from students and faculty or through regular staff meetings where educators share best practices, challenges, and potential solutions. This continuous process of monitoring and reflection enables institutions to revise and enhance their policies in a timely and responsive manner, thereby ensuring alignment with the dynamic and evolving landscape of AI technologies.

4.3. Strategic use of AI in teaching

When addressing the challenges posed by AI, it is not sufficient to merely restrict its use. Universities must also learn to use AI to fully catch the opportunities and potential it offers for the development of higher education.

As Fan Ouyang, Pengcheng Jiao have noted, there are two paradigms of Artificial Intelligence in education. The initial paradigm, termed AI-supported learning, is defined by a partnership between learners and artificial intelligence tools. In this paradigm, AI serves as a supportive tool by collecting data on learners to adaptively optimize the student model. Learners, in collaboration with AI systems, focus on their learning process. And during this process, the AI system and students should establish active interaction with each other to achieve better and more effective learning [12].

The second, more practical paradigm is AI-empowered learning, in which the learner acts as the leader. In this model, education is treated as a complex adaptive system where collaborative efforts among learners, instructors, and AI technologies are essential to enhancing learners' intelligence effectively helps learners and instructors enhance their intelligence by providing a high level of transparency and accuracy, such as promoting learner-centered learning through explanations and personalized support. Or learners can take AI as the leader of their learning, manage the risks of AI-determined automation, and thereby enhance learning efficiency.

Overall, universities and colleges should encourage AI as a supplementary tool for teaching in higher education rather than a replacement for human educators. By leveraging technology appropriately, institutions can expand access to learning resources, improve teaching standards, and be better positioned to seize the educational development benefits brought by the advancement of AI technology.

5. Conclusion

In conclusion, AI is a double-edged sword for education. Its emergence has brought significant developmental benefits to higher education. AI-driven feedback and intelligent tutoring systems both enhance student engagement and promote higher learning outcomes through greater personalization and efficiency. Additionally, the integration of big data enables universities to make more informed decisions and allocate resources more effectively.

However, the adoption of AI also introduces new challenges. Concerns regarding the quality of academic learning, the accuracy of AI tools, threats to academic integrity, and the risk of deepening educational inequality all require careful institutional responses.

To effectively leverage the potential of AI without exacerbating its drawbacks, universities must adopt thoughtful and strategic implementation practices. These include designing AI- Developing AI-resistant assessments through diversified and personalized evaluation methods, establishing clear ethical guidelines informed by timely feedback, and integrating AI into teaching practices in ways that promote its responsible, equitable, and effective use in higher education.

Ultimately, the emergence and application of AI are not intended to completely replace the existing human teaching methods. Instead, colleges and universities should embrace the development potential of AI while proactively developing strategies and frameworks to address the associated risks and challenges.