AI in Academia: Navigating Unprecedented Transformation

The advent of Generative AI tools has sparked discussions about an unparalleled transformation across various domains, including academia. Organizations are grappling with the application of AI in educational contexts.

At Lund University, a prominent Swedish institution, educators wield the authority to grant students access to artificial intelligence for assistance with their assignments.

In Perth, at the University of Western Australia, faculty members engage students in dialogues concerning the potential challenges and advantages of employing Generative AI in their academic pursuits. Similarly, the University of Hong Kong has cautiously embraced the use of ChatGPT within defined boundaries.

ChatGPT, launched by OpenAI with support from Microsoft on November 30, has rapidly ascended to become the world's most rapidly adopted application, spawning competitors like Google's Bard.

Generative AI, including ChatGPT, relies on language patterns and data correlations to generate a wide spectrum of content, ranging from essays to videos and even mathematical computations that superficially resemble human output. This phenomenon has led to discussions about a revolutionary transformation in numerous sectors, academia being a prominent one.

Academics are among those who face a potential existential threat if AI can replicate research tasks currently undertaken by humans, and at significantly accelerated speeds. However, many also recognize the benefits of Generative AI's capacity to process and analyze information and data, thereby providing a foundation for more profound critical analysis by humans.

Leif Kari, the Vice President for Education at KTH Royal Institute of Technology in Stockholm, suggests, "It can assist students in tailoring course materials to their individual needs, much like a personal tutor would."

The United Nations Educational, Scientific, and Cultural Organization (UNESCO) introduced global guidance on the use of Generative AI in education and academic research on September 7th, marking a significant milestone.

For national regulatory bodies, this guidance delineates steps to be taken, including data protection measures and the revision of copyright laws, and advocates for the development of AI competencies among educators.

Some educators draw parallels between AI and the introduction of handheld calculators into classrooms in the 1970s, which initially sparked debates about their impact on learning before ultimately being embraced as indispensable tools.

However, concerns have arisen that students might overly rely on AI for producing work, potentially leading to academic dishonesty. The improving quality of AI-generated content over time further raises concerns. Presenting GenAI-generated work as original may also pose copyright issues, prompting debates about whether AI should be prohibited in academic settings.

Rachel Forsyth, a Project Manager at Lund University, acknowledges the difficulty of enforcing a ban, emphasizing the importance of shifting the focus back to learning rather than policing students.

For decades, Turnitin has been a primary tool for detecting plagiarism globally. In April, it introduced an AI-powered tool designed to identify AI-generated content. Although this tool has been made available free of charge to over 10,000 educational institutions worldwide, Turnitin plans to introduce fees starting in January.

According to Turnitin data, only 3% of students have used AI for over 80% of their submissions, while a majority (78%) have not used AI at all. Issues have arisen in the form of false positives, where text authored by humans, including professors testing the software, has been erroneously identified as AI-generated. However, those falsely accused of using AI can defend themselves by retaining multiple drafts of their work.

Students themselves have been experimenting with AI, and opinions vary. Some view it as capable of basic summarization but emphasize the need for fact-checking, as GenAI cannot discern fact from fiction or right from wrong. Additionally, its knowledge is limited to internet-scraped information, which may not suffice for highly specific queries.

Sophie Constant, a 19-year-old law student at the University of Oxford, remarks, "I believe AI has a long way to go before it becomes truly useful. I cannot rely on it for a single case, as it lacks access to the articles I am studying, making it less helpful."

UNESCO's latest guidance highlights the risk of deepening societal divisions due to unequal access to electricity, computers, and the internet as educational and economic success become increasingly dependent on these resources.

Efforts to regulate AI are underway, with the European Union (EU) at the forefront, although its draft legislation has yet to be enacted into law. While the regulations do not specifically address AI in education, their broader ethical principles could be applied to the field.

Post-Brexit, the United Kingdom is also working on guidelines for the use of AI in education, consulting with educators and planning to release results later this year.

Singapore, a leader in training teachers in AI utilization, is one of nearly 70 countries that have developed or planned AI strategies.

Kirsten Rulf, a Partner at Boston Consulting Group who previously played a key role in negotiating the European Union AI Act, advises, "Rather than resisting it, as a professor, you need to harness AI, gain experience with it, establish a robust framework, guidelines, and a responsible AI system, and collaborate with students to find an effective approach that suits your needs."

Rulf's previous role as Head of Digital Policy at the German Federal Chancellery contributed to shaping the European Union's AI regulations.