“Generative AI”, “Large Language Models” (LLM), and “Chat GPT” have become part of our everyday vocabulary in the last few months. The catalyst was the launch of “GPT 3.5”. The platform became publically available in November 2022. This has been a unique LLM as it came with a super easy chatbot interface which is conversational AI. Further, there has also been a follow up sequence of various revolutionary tools which are having a massive impact on our systems and structures such as Bard, Cohere generate, DALL-E 2 et c. Experts say we are already in the Fourth Industrial Revolution facilitated with the power of AI along with the Internet of Things, 3D Printing, Robotics, and others.

In this blog the focus is to understand the impact of Generative AI in the higher education sector, that too related to teaching and learning. From students’ perspective, Generative AI can provide them assistance with their assignments/ projects. Hence, there needs to be guidelines for its appropriate use and also a strong emphasis on checking the factuality of the contents generated using these tools. An interesting question is, how can such generative algorithms facilitate better learning for students, especially in the higher education sector?

On the other hand, the teachers have questions regarding how to check if students are “cheating”; how to learn these tools and integrate it into their pedagogy; whether will there be an over-reliance on these tools, and again what can the students learn? Research is the most important part of academia in general. Even taught Master’s degrees involve projects in the final evaluation, and with them comes the question of ethics and integrity. Plagiarism is considered unacceptable in educational institutes. There are standard sets of guidelines for citing and referencing which leads the readers and reviewers easily to the source. They establish the credibility of the research work and add the produced work to the knowledge base.

The biggest issue with these models is that they are not mature enough to be error-proof. For example, one of my colleagues uploaded an ‘unpublished article’ on GPT 3.5 which he wrote in 2017 on Chat GPT 3.5 to check if the article was written using AI. The tool claimed that it had written the article. It is interesting to note that in 2017 this tool did not even exist. On probing again it “apologizes” and claims it is written by a human. Then again it says it is AI-generated. The point to understand is that it is a probabilistic model. It doesn’t give definite answers (see image 1).

(Image 1)

Similarly, I once put a command asking “give me the top 5 good articles on education policy” on GPT 3.5. I received 5 articles, their heading, author names, and hyperlinks. Surprisingly, these hyperlinks lead to “error 404 page not found” or the link was not working. When I looked up the article names along with the author and publisher names on a simple google search these articles did not exist. (see image 2)

(Image 2)

Experts say that we do not know the full capability of generative AI models. There are both optimist and pessimist views on this new technology as has been the case with the advent of any new technology (for example when calculators, or computers etc. were invented). The past shows us we need to learn to adapt to new technologies.

What we need to do further is to teach students from all streams the basics of research, modeling, and thinking critically. AI is disrupting many areas of work which include a variety of skills, be it multimedia work such as photo–video editing like giving automatic subtitles or voiceovers in any language, virtual assistants such as managing calendars, writing mails, preparing documents and report formats, programming, creating websites, arts, music and many more. These are the skill sets that are becoming easier by learning “prompt engineering“, basically giving better prompts to the Generative AI models to get optimal results efficiently. Learning the content is easier but more than this designing opportunities for students to teach how to learn and assessing the same is required in higher education institutes with the greater influence of Generative AI.

Note: This blog draws inspiration from the presentations of Dr. Manu Kapur and Dr. Pascale Fung during “Dialogues on Asian Universities” on “AI and Higher Education: Implication, Challenges and Opportunities”.

*The Kautilya School of Public Policy (KSPP) takes no institutional positions. The views and opinions expressed in this article are solely those of the author(s) and do not reflect the views or positions of KSPP.