Artificial intelligence (AI) is rapidly moving from science fiction to everyday reality, transforming the way industries work, and higher education is a big part of that change. Universities around the world are now implementing AI in classrooms, research, and day-to-day management.
According to a 2023 UNESCO report, almost 89% of universities were already involved in some kind of AI project by the middle of the year. The World Economic Forum has even estimated that AI’s contribution to the global economy in 2030 will be approximately $15.7 trillion, and education will be a major sector of this.
For all the latest news, follow The Daily Star’s Google News channel. In the US, the AI market in education is valued at $4 billion in 2022 and is projected to grow to $20 billion by 2027 (HolonIQ). The application of AI in education is of great importance. Personalized learning, frictionless management, thriving research, and more, but it also raises the need for serious discussions about ethics, equity, and the overall quality of education. This article explores both sides of the debate: the ways in which AI is proving useful in higher education, and the areas in which it continues to have challenges.
AI’s ability to personalize learning is one of its key capabilities. AI-powered platforms can monitor learner progress in real-time and adapt lessons to match their learning speed, rather than applying a generic approach.
A study of 12,000 students at Arizona State University (ASU) found that the use of AI-assisted adaptive tutoring increased pass rates by 11% and reduced the number of students dropping out of courses by 14% (ASU EdPlus, 2022). Global demand for adaptive learning tools doubled from $1.3 billion in 2020 to $3.1 billion in 2023 (MarketsandMarkets).
AI has made it possible to close the educational gap in third world countries. For example, India’s NPTEL (National Program on Technology Enhanced Learning) project uses AI to recommend courses to more than 1.5 million students every year. A 2024 study found that the use of AI tutoring increased course completion rates in online learning for students from disadvantaged backgrounds from 5% to 18%.
Universities spend a huge amount of their time managing everything from responding to student questions to processing admissions. AI is now helping ease that burden. Georgia State University’s AI chatbot responds to more than 200,000 student inquiries each year. According to Educause Review, 2023 sources, implementing this simple innovation reduced summer “melts” (students who are accepted but do not enroll) by 22%, resulting in universities saving nearly $1 million on an annual basis.
Artificial intelligence is also being applied to predict the likelihood of students dropping out. The University of Maryland devised a system that could identify at-risk students with 85% accuracy, which ultimately increased retention rates by 5% (WEF, 2024). Meanwhile, the University of Michigan applies machine learning to evaluate more than 70,000 applications each year, accelerating processing times by 40%. According to McKinsey forecasts (2023), the adoption of AI-based automation in higher education could ultimately result in total savings of $100 billion for universities around the world by 2030.
AI has become a powerful research partner in universities. Tools like IBM Watson and Google Scholar’s AI engine can scan thousands of academic papers in seconds, helping researchers find patterns and gaps in existing knowledge. A Nature study (2024) showed that an AI-powered literature review can reduce research time by 60% for PhD students. In the scientific field, AI is currently being used to simulate laboratory experiments. For example, MIT’s AI drug discovery system identified new compounds 50 times faster than traditional methods.
Collaborative programs such as the EU’s Horizon Europe initiative have also shown results, with AI-supported research teams producing 15% more publications per euro invested than those using traditional methods (European Commission, 2025).
AI also improves access for students with disabilities and limited resources. Microsoft’s Immersive Reader, integrated into systems such as Canvas, provides real-time translation, text-to-speech, and visual support, helping 15% of college students with disabilities in the United States (NCES, 2023).
In Bangladesh, the 10 Minute School app delivered AI-powered micro-lessons to more than 500,000 advanced learners in areas with limited internet access. Similarly, Open University UK found that using voice-assisted AI increased participation by visually impaired students by 28%.
AI systems rely heavily on student data, which raises legitimate concerns about privacy and security. A 2024 Ponemon Institute survey found that 62% of students are concerned about how their data will be used. Europe’s GDPR sets strong privacy standards, but fewer than half of EU universities are fully compliant. In 2023, a major university data breach exposed 1.2 million student records, showing how vulnerable education data can be.
AI tools are only as fair as the data used to train them. Unfortunately, algorithmic bias can lead to unfair outcomes. A ProPublica (2022) investigation found that some AI predictive systems classify minority students as “high risk” more often than other students, increasing dropout rates for those groups. A Stanford University (2023) study found that AI systems tend to underrepresent women and minority students in STEM courses, further widening the gender gap.
Infrastructure is another issue. Only 40% of universities in low-income countries have reliable internet for AI use, compared to 95% of universities in high-income countries (World Bank, 2024). In Bangladesh, higher education enrollment remains at around 20% (UNESCO, 2023) and poor connectivity limits the effective use of AI.
Over-reliance on AI is detrimental to true learning. An OECD (2024) study found that students who used AI to write their essays scored 15% lower in analytical skills. Meanwhile, plagiarism detection software such as Turnitin reports detecting AI-generated text in 10% of student papers in 2023, up from just 1% two years ago.
Teachers’ opinions also varied widely. According to a Times Higher Education (2024) poll, 58% of professors are concerned that AI will take over some of their roles. Although AI systems for grading are faster, they continue to struggle with tasks that require subjective judgment, especially in the arts and humanities, where error rates can be as high as 20% (Inside Higher Ed, 2023).
Setting up AI infrastructure is expensive. A mid-sized university can spend between $500,000 and $2 million just to launch an AI-enabled learning system (Gartner, 2024). Teacher training adds an additional layer. A UK study found that upskilling teachers for AI integration costs around £5,000 per teacher. According to the Asian Development Bank (2023), only one in four institutions in Asia currently has the budget to fully implement AI.
Several success stories stand out. Carnegie Mellon University (USA) uses AI tutors to improve computer science grades by 12%. At Strathmore University (Kenya), AI-based instruction increased graduation rates by 9%.
However, not all experiments went smoothly. In India, AI-powered online exam monitoring during the coronavirus pandemic incorrectly flagged 15% of students for cheating due to poor lighting and internet issues, leading to a legal dispute. Examples like this remind us that technology alone is not enough, context matters. Similarly, at South Korea’s Yonsei University, approximately 600 students in the “Natural Language Processing and ChatGPT” course were suspected of cheating using AI tools such as ChatGPT during online midterm exams, despite strict monitoring. A student poll found that 190 out of 353 admitted to using unfair means, suggesting more than half of the class was involved. This case highlights that misuse of AI is increasing faster than organizations are prepared. Even though 91.7% of Korean university students use AI in their classes, 71.1% of universities lack clear policies (Korea JoongAng Ilbo, November 9, 2025).
AI has the power to transform higher education by improving learning outcomes by nearly 18%, saving billions of dollars, and making education more accessible to millions of people. But it also poses serious challenges related to stigma, privacy, and affordability. To move forward in a responsible manner, universities must first establish ethical guidelines and then invest in teacher training and infrastructure. The IEEE AI Ethics Guidelines (which 40% of top universities now follow) emphasize the importance of transparency, accountability, and fairness among its key principles.
AI’s role in education should be to support human teachers, not replace them. By taking precautions and implementing balanced integration, higher education systems can harness the potential of AI to prepare students for the future, a future in which 85 million jobs will be relocated or transformed by 2030 (WEF). The goal should not just be smarter technology, but smarter, more inclusive learning for everyone.
Dr. Carmen Z Lamagna He is a member, director and official vice president of the American International University of Bangladesh (AIUB).
Professor Dip Nandi Dr. He is the Associate Dean of the Faculty of Science and Technology at the American International University of Bangladesh (AIUB).
