(Crowdfunded Paper) Higher Education in the Age of A.I.: An Exploratory Narrative Review
TipTipAuthor: Edmilson Rodrigues do Nascimento Junior (MsC in Computer Science at CIn/UFPE - ernj@cin.ufpe.br)
Abstract
The advent of generative A.I. tools like ChatGPT and Bard created shockwaves in universities around the world. Now, anyone with internet access can use A.I. to generate or comprehend texts with a higher education level competency. What will be the role of higher education institutions in a society transformed by A.I.? This paper conducts an exploratory narrative review on papers with relevant keyword selection. It discusses about the evolving roles of universities in society, the impact of A.I. in the labor market, how to adapt the education systems to the age of A.I., the skills for humans to stay relevant in the age of A.I. and the challenges and opportunities for the adoption of A.I. in universities. The times have always been changing, but now, it is changing faster than ever. Therefore, higher education institutions need to adapt.
Keywords: Generative Artificial intelligence, Higher Education, Academia, chat-gpt, education 4.0
Introduction
ChatGPT is a Generative Artificial Intelligence (GAI) based on a Large Language Model (LLM) that was launched by the startup Open AI on November 30th 2022 and caught the world's attention. In little over a year, it has grown to 180.5 million users globally, of which, 100 million are active users (107 Up-to-Date ChatGPT Statistics, accessed in January 2024) and use it at least once per week. Microsoft invested in OpenAI, and other major players started to develop their own LLMs.
However, ChatGPT is just one application of Artificial Intelligence (A.I.), a technological trend that is transforming society and the way people work and learn. A McKinsey Global Institute study (Ellingrud, 2023) estimates that 800 million jobs globally will be displaced by A.I. until 2030. Many other jobs will be created in the process. But, are universities today preparing their students for the jobs of the future?
Some authors (Ouyang and Jiao, 2021) claim that the education system today could better prepare its graduates to live and work in the age of A.I. The authors start by recognizing that there are tasks that computers can perform better, like recalling data, recognizing patterns and statistical reasoning. And there are tasks that humans can perform better, like empathy, self direction, common sense and value judgements. Therefore, the pedagogy of the era of A.I. puts emphasis on human skills, like critical thinking, communication, collaboration and creativity, and also to collaborate with ubiquitous A.I. tools that will transform life, learning and work. In fact, in the same way that today's professionals require “computer literacy” to work in the age of the ubiquitous internet, tomorrow´s workers will have to have “A.I. literacy” so they can build their agency, employability and their ability to contribute to society. Professionals in the age of A.I. will have to be life-long learners and understand how the A.I. technologies work, in particular how data is selected, manipulated and interpreted.
Also, A.I. is transforming education itself and it is developing in multiple directions, may it be student-facing A.I (for support to learning and assessment), teacher-facing A.I. (to support teaching) and system-facing A.I. (to help in the management of educational institutions).
Additionally, Ouyang and Jiao (2021) defend that the use of technology must be aligned with the correct pedagogical perspectives. And they do that by proposing three AIEd paradigms: 1) AI-Directed, learner-as-recipient. 2) AI-supported, learner-as-collaborator, 3) AI-empowered, learner as leader.
In the first paradigm, the AI represents the domain knowledge and directs the learning process. On the other hand, the learner is the recipient of such knoweldge and follows a specific learning pathway. This is based on the theory of behaviorism, that sees learning as a reinforcement of knowledge acquisition through programmed instructions. It offers new concepts in a logical and incremental way and offers the users immediate feedback about responses and tries to maximize positive reinforcements. Then, the learner is the recipient of pre-defined sequences of knowledge. The A.I. system does not model itself to the learner's incoming knowledge and skills, nor adjust its feedback to the learner. It is the paradigm that is least learner centered. An example of implementation of AIEd in paradigm one is ITSs (Intelligent Tutoring Systems).
In the second paradigm, the authors see the A.I. as a supporting tool, while the learner is the collaborator and the system has a focus on the individual's unique learning process. The theoretical theory which supports this paradigm is constructivism, that defends that learning occurs as individuals interact with people and that information and technology are socially situated constructs. Therefore, for this paradigm, the A.I. system and the learner should interact and work as a team to optimize the learner-centered, personalized learning experience. Some A.I. implementations that use paradigm two are the dialogue-based tutoring systems (DTSs) or the exploratory learning environments (ELEs).
In the third paradigm, the authors view education as a complex adaptive system that requires a synergy between many entities (student, instructor, information and technology). The authors claim that this paradigm recognizes that the instructor has to be equipped with support to foster learner-centered learning. Therefore, it is up to the learner to take the agency of her own learning and manage the risks of the AI decision. Then, the ultimate goal of the application is to augment human intelligence and potential, but the human remains the leader of this process. Then, paradigm three defends the synergetic collaboration between the A.I. and human intelligence to produce adaptive and personalized learning.
Much has been discussed about how A.I. will impact education and society, but there is a gap in literature regarding how higher education institutions should adapt and better prepare their graduates for the era in which A.I. is pervasive. Therefore, this paper has the objective to understand the narratives of the available literature regarding how artificial intelligence is impacting higher education institutions and how universities should change to better prepare their graduates for the future of work. It does so by conducting an exploratory narrative review on papers with a relevant keyword selection.
Methodology
Being such a transformative technology, A.I. is expected to have a large impact on higher education institutions. This study, then, aims to conduct an Exploratory Narrative Review on the topic to understand the main themes in the literature regarding the role of universities in the era of A.I.
The following keyword selection was employed: “higher education in the age of A.I.” and “Impact of AI on higher education” in the Google Scholar database and only papers published in 2019 or more recently were chosen. Then, fourteen documents were selected based on the suitability of title and description, and three of them were discarded because they were not open access. Therefore, eleven papers were included in this study.
We were guided by the following research question: What is the current narrative of the academic literature, from 2019 to 2023, on the impact of A.I. on universities and what will be the role of higher education institutions in the age of A.I.?
We could then distribute the topics of the papers in the following themes:
- The evolving roles of universities in society
- The impact of A.I. in the labor market
- How to adapt the education systems to the age of A.I.
- Skills for humans to stay relevant in the age of A.I.
- Challenges and opportunities for the adoption of A.I. in universities
What are the evolving roles of universities in society?
Salmon (2023) makes an interesting reflection about the evolving role of universities in society. She makes a parallel of the evolution of universities with the evolution of the internet itself and classified the evolving role of education in the last 1000 years according to figure 1:
(from Salmon, 2023)
The education 1.0 is the era of “transmission”, when the first universities in Europe appeared in Bologna, Paris, Oxford and Cambridge. That was a place where academics and their helpers would be supplied with information and be able to learn from practice and from each other. Then, they would come out as “university educated people”.
Around 60 years ago, the notion of “going to the university” started being challenged as more people wanted access to the university and the concept of “Open University” (OU) emerged. Education 1.25 , then, allowed for open entry and access and flexibility for work and domestic responsibilities. Education 1.5 innovated true distance learning, allowing students to take classes with learning materials at home and doing exams in test centers. And around 20 years ago, technology allowed for the creation of “Virtual Learning Environments”, enhancing education via the web. While it allowed for the inclusion of millions of people in the higher education environment, it was all based on the “knowledge transmission” paradigm.
The education 2.0, then, emerged around 2005, when the web 2.0 or the “read-write” web came about. Things like blogs, wikis, images and videos started being used by academics and the virtual learning environments themselves started incorporating it in their systems. Open education repositories and crowd-contributed content started emerging, like wikipedia. Then, education 2.5 started to emerge. It was a challenge to the concept of “transmission of information”, and concepts like “blended learning” and “flipped classroom” started emerging. New concepts of online higher education emerged, like the MOOCs (Massively Open Online Courses).
Afterwards, education 3.0 emerges together with the “Semantic Web”. Web 3.0 or the Semantic Web refers to the paradigm where we take the internet together with us everywhere, in other words, it is ubiquitous. Because of that there is so much data available that applications can start “making sense” of it or to become smarter. Because of this, the author suggests that universities are no longer the “source of truth and learning”. And, because of that transmission alone is insufficient. The author mentions that most students entering universities now have been born in the internet age where they have had access to the internet their entire lives.
The age of education 3.0 has also questioned many assumptions about higher education. For instance, it highlighted the importance of access and diversity for students, staff and teachers. Also, the relevance of topics and the employability of students came into question.
Additionally, Salmon reflects about the different Industrial Revolutions and its impacts on society. The first industrial revolution around 1760 used water and steam and impacted transportation, urbanization and factories. The second industrial revolution used electricity and created mass production. The impact of mass industrial process reflected on education, with standardization, rote learning and very large classes. It was in 1969 that Peter Drucker coined the term “Information Age” and “knowledge economy”, making explicit the link between Research, Higher education and the success of a nation. Therefore, the third industrial revolution used electronics and information technology, and automated production and created global supply chains. The advent of computers and the internet disrupted nearly all industries, but most importantly, the way that people teach and learn.
The author claims that, around the year 2010, there was a big shift in universities to emphasize employability, the students´ skills and be centered on the student experience. That is what she claims to be the big focus of education 4.0.
The rise of education 4.0 will be made possible thanks to the web 4.0, that is enabled because of the Web 4.0. Called “The Symbiotic Web”, web 4.0 receives this name because of the symbiosis between the human and artificial intelligences and how they interact to become a “team”. When everyone has an intelligent agent in their phones or search engines, this is already impacting the way people learn and solve problems.
Salmon points out that Artificial Intelligence is becoming competent in tasks that only highly skilled humans could perform. This is giving rise to the 4th Industrial Revolution. And the education 4.0 together with the industry 4.0, is giving rise to the Globalization 4.0.
And what is the role of education in this brave new world? The author suggests:
“Implications for education may involve quite dramatic variations in the demand for knowledge and skills as well as expanding possibilities for teaching and learning (...). Babies born today not only will never know a world without the Internet, but one in three may live to be over 100, some to 120 or more (...). So, there is also a very strong case for a sustainable life-long love for learning and personal transformation (...)”
What will be the impact of A.I. in the labor market?
Professor Webb (2019) created a method to predict the impact of a technology in the labor market. His method identifies which tasks can be automated by the introduction of a particular technology (such as A.I.). That way, identifying the level of exposure of different occupations to such technology. Then, he empirically applied the method to two past technological advances that had significant impacts in the labor market: automation (robotics) and software, and determined the relationship between the exposure scores and the evolution of wages that the advent of automation and software brought upon.
Webb (2019), then, applied the method to Artificial Intelligence. The data suggests that the exposure that will be brought upon on the labor market by A.I. will be highest for high-skilled occupations. In addition, his study suggests some of the occupations that are likely to be affected by the ever increasing power of A.I. are: Clinical laboratory technicians that conduct visual and analytical work on tests, chemical engineers that design and operate production processes, Optometrists that detect diseases in the eye, power plants and data center safety operators that need to control the safety of energy outputs. On the other hand, some low-skilled workers are also at risk of being exposed to the advent of A.I., for instance, jobs that require inspection and quality control.
The occupations that are least exposed to the advent of A.I. on the labor market are those that involve interpersonal skills, like professors, managers, care-takers, or that require fine motor skills, like baristas, food preparation workers and massage therapists.
The paper concludes that A.I. is very likely to affect different occupations, and that are high-skilled workers that are more exposed. The data also suggests that while there will be a decrease in the inequality between the 10th and 90th percentile of wages in a profession, there might be a big increase in inequality between the 90th and 99th percentile in the same profession. The paper wraps up by recognizing that the method can measure the impact of A.I. in the demand of the labor market, it does not precisely assess how the same technology may impact the labor supply, for instance, by being used for the training of new workers.
Additionally, the McKinsey Global Institute study (Ellingrud, 2023) discusses which industries are more likely and least likely to be displaced by automation. While focused only on the US job market, it is a good indicator of the state of the future global job market.
In the graph below, the X axis represents the increase in the automation adoption brought upon by A.I., and the Y axis represents the change in the labor market from the years 2022 to 2030. The size of the bubbles represent the size of the impact. If we divide the graph in four quadrants, the workers in the industry that are in the upper right quadrant are going to benefit from an increased demand in labor and are going to have their work activities impacted by generative A.I., while the workers in the industry in the lower left quadrant will suffer a decrease in labor demand and will have a modest change in their work activities.
(from Ellingrud, 2023. Page 7)
In a world with widely available Artificial Intelligence, what knowledge, skills and attitudes do people have to have so that they are not easily replaceable by A.I. ?
The current scenario begs the question: What should we teach people today for them to remain relevant and employable in 2030?
Rampersad (2020) defends that the workers of the future will require a different set of skills. These are sometimes referred to as generic skills, professional skills or “soft skills”. For instance, critical thinking, communication, team work, innovation and to have an entrepreneurial mindset.
In addition, the author claims that Innovation will be one of the most important skills in the era of industry 4.0, and making the transition to the industries of the future will not only be a technological challenge, but also a human issue. Human factors are critical elements so that the workers of the future can survive and adapt to new technologies and changing workplaces. Therefore, a skill set transition is essential for the workers of the future.
Prof. Rampersad describes innovation as ‘the process of bringing into being something novel and useful’. It involves all the activities from creative thinking, to idea generation and the creation and commercialization of new products. There is extensive literature that suggests that some of the most important skills that will be required by the innovators of the 21st century are 1) problem solving, 2) Critical thinking, 3) Communication and 4) Team Work.
The author (Rampersad, 2020) defines (1) Problem Solving as “an ability to analyse and transform information as a basis for making decisions and progress towards the solution of practical problems”. It is a very important skill for the worker of the 21st century who will need to develop solutions for complex technological, economic and social problems. The author also says that (2) Critical Thinking is the capacity for “logical, analytical, conceptual and reflective reasoning”. Despite being widely recognized as one of the most important goals of higher education, many authors recognize that graduate students still lack this important skill. In addition, the paper mentions that (3) Communication is an important factor for innovation. It is defined by the author as “the ability to use language, symbols and text interactively” and it includes both verbal and written communication. Finally, the author say that (4) Teamwork is an important skill and enabler of innovation and defines it as “the ability to work constructively with others on a task”
Some approaches widely recognized as to incentivize this important skill in students are the creation of “makerspaces” to foster new ideas, the incorporation of teaching about innovation, entrepreneurship and design. It is important to highlight that the teaching of entrepreneurship should not only be focused on creating new ventures, but also intrapreneurs, to cooperate with existing firms.
Another set of skills that the workers of the future must have is AI literacy. Some authors (Ng et al, 2021) claim that, in the same way that today's workers must have digital literacy to be competitive in the job market, tomorrow´s workers will be required to have A.I. literacy skills, such as the ability to understand the basics on how the technology works, to distinguish between its ethical and anti-ethical practice and use A.I.´s full capabilities in advanced ways. Therefore, the authors defend that these skills should be taught now only in universities, but on K-12 levels. According to the authors, it should be included in K-12 STEAM education packages via playful learning experiences.
The authors (Ng et al, 2021) claim that A.I. Literacy encompasses the ability to the individual to (a) Know and Understand A.I., that means “to apply A.I. knowledge, concepts and applications in different scenarios” -; (b) Use and Apply A.I., that refers to the ability to use A.I. thinking to solve problems, handle data and interpret findings; and (c) Evaluate and Create A.I., that means the ability to “critically evaluate A.I. technologies, communicate and collaborate effectively with it.”
How can we adapt our higher education institutions to the age of A.I. ?
It is widely recognized that higher education institutions need to change. But what needs to change? And change to what?
Williams (2019) reflects that era of A.I. require universities to change and adapt. The author recognizes that the curriculum models of universities in which, at the end of a journey concedes a degree to students, and that is an organization divided into structures like faculties, departments and disciplines, is the result of a system in which the boundaries between professions and knowledge was clearly delimited, and the academic content was valid for decades. However, in the era of A.I., there are different forces changing work and life in the 21st century. For instance, the rise of the knowledge-intensive sector, the reduction of the “half-life” of knowledge and the syncretism between disciplines. These facts call for a rethinking of the academic model from the 3-4 years delivery model to a “just-in case”, with flexible content and lifelong learning models, what some refer to as “just-in-time learning”.
Additionally, Salmon (2019) reflects about what needs to change so that universities achieve Education 4.0. She defends that higher education institutions need to create new Curricula, that uses data and extrapolate the students´ interests and “future-proof” it by creating new curriculum portfolios; new Modes of Learning, that allow people to teach and learn collaboratively by design. Assessment and credentialing must also change; And there needs to be a Rethinking of achieving, where universities should open their minds, and their doors to new ways of thinking. For instance, working with commercial companies. Finally, the author concludes stating that universities do not need to change their missions and values. But they need to change their methods in order to educate the workers of the future.
Complementarily, Williams (2019) defends that higher education institutions must guide their students towards employability 4.0. In the era of A.I., students will have to have skills like “deeper relationship building, flexible and innovative thinking, social and emotional intelligence, collaborating virtually, design mind-sets, new media competencies, ‘thinking like a data-scientist’ and interdisciplinarity.” Another hint that the author gives is preparing students towards the great challenges of our era, like environmental sustainability, income inequality and the expansion of cities. In addition, Salmon (2023) points out that the jobs of the future will require “interpersonal skills, higher order cognitive development and preparing to work in globalized contexts, (...) the important role of ‘Active Learning’ in order for students to be able to rapidly assimilate and work with change, problem solving, judgment and decision making”.
According to Williams (2019), one methodology that is well suited for the education of 21st century students is Competency Based Education (CBE). It emphasizes that students learn explicit and transferable learning objectives in an environment that provides timely and personalized support and formative feedback. That way, the students can apply a different set of skills and foster critical thinking, problem-solving, communication and collaboration.
The 21st century job market also requires a new curriculum content and delivery. The author suggests that universities need to change focus from academic disciplines to a “procedural knowledge-action, nurturing soft-skills and developing students as whole persons, making them adaptable to changing circumstances”. The author affirms that universities should assume new missions. In addition to course delivery, they should be “education service advisers”, helping the students to develop a graduate level portfolio of educational assets, may it be course credits gained in-house, in other prestigious institutions around the world or through extracurricular activities.
What are the Challenges and Opportunities for the adoption of A.I. in universities?
Some challenges of the application of A.I. in the university setting is the importance of data ethics and algorithmic biases. It continues by mentioning the impact that A.I. will have on the teacher role (allowing them to focus on the human aspect of teaching, rather than the burdens of monitoring progress of the students). In addition, the use of A.I. is expected to impact the learners´ agency, like their resourcefulness, critical thinking, independent thought and other 21st century skills.
(Grassini, 2023) One big limitation that the LLMs currently have is that sometimes they are not accurate or reliable in the information presented, what is called by some a “hallucination”. Another limitation is that if there is a lack of quality and diversity of data used in the training, the end result might present biases. Finally, another issue is the possibility of student plagiarism. Currently there is an arms race between plagiarism detectors and the A.I. technology creating it. Thus, it is very important to indoctrinate the students on the importance of academic honesty and the value of the originality of the work presented. It is up to the universities and regulator authorities to provide clear guidelines of “fair use” of Chat GPT and other tools.
On the other hand, regarding opportunities, Grassini (2023) mentions different instances on how A.I. can be used to enhance the learning and pedagogical experiences. For instance, to automate and improve the grading system, and auto-suggest the strengths and weaknesses of the student's work. Therefore, the educator can be free to spend more time providing feedback and help to the pupils. However, one possible limitation for this use case is that A.I. require a big amount of prior data, and thus it may be more suited for standardized tests and nationwide professional educational examinations, like the SAT, LSAT, Bar exam, etc.
Another use case for A.I. in education is the translation of educational materials in different languages and the creation of interactive and adaptive learning environments. Through individualized A.I. tutors, chatGPT and other systems can adapt to the individual needs, learning styles and progress of students, what is known as Adaptive Learning. Finally, chatGPT and similar systems offer the opportunity for the educators to enhance their pedagogical practices, creating presentations, lesson plans and course materials.
Conclusion
It is important to recognize that A.I. tools are here to stay (Grassini, 2023), and it is more realistic to accept its presence in the academic world and integrate in the curriculum. Banning it would be similar to banning calculators in math classes or banning search in homework, and would put the students that do not learn to work with such tools at a disadvantage.
Education is widely recognized as a field that is traditionally conservative (Lim et al, 2023). But the speed in which this technology is developing has caught everyone by surprise to the point of creating four paradoxes:
The first paradox is that A.I. is at the same time a friend and a foe for the education system. It can be used to facilitate the writing of texts, code, essays and scripts, and the correction and timely feedback on a scale while the students are executing such tasks. But it can also be a foe for the education system because it poses the challenge for educators that the students might simply copy and paste the answers.
The second paradox is that A.I. technology is at the same time capable and dependent. It can generate responses in different fields of knowledge. However, the responses it generates are completely dependent on the inputs it receives. Thus, at least the version until January 30th 2023, was not able to correctly choose references to support its answers.
The third paradox is that generative A.I. is accessible yet restrictive. It is accessible because it is at the core of ChatGPT´s mission to ensure its availability to all of humanity. It can be used by professors around the world to curate and translate knowledge for their students, for example. However, there is the risk of the companies making available for a period of time only, or to release newer versions under a paid subscription service. This would exclude those who can not afford the premium fee for having access to the better version and would widen the social-economic gap.
The fourth paradox is that generative A.I. gets more popular when it gets banned. Banning the use of ChatGPT and similar tools was the standard response from education bodies around the world at first glance. But this can have the opposite effect, because the attempt to limit the technology can lead to increased awareness and interest from users.
Therefore, while some see Artificial Intelligence as a Ragnarok - the destruction of the education system -, others see it as an opportunity for accessible and high quality information for all. The time has always been changing, but because of the new developments with A.I., it is changing faster. So, it is paramount that universities and educational bodies update their policies and guidelines regarding plagiarism and to what extent it is possible to use generative A.I. in the assignments. It is also important to embrace a culture of change to allow the students to be up to date with the future technologies. Banning it will only exacerbate existing inequalities in the system.
This study conducted an exploratory narrative review of the role of higher education in the age of A.I. . It had the limitation of using only choosing papers from 2019 to 2023, that are open access and available on the Google Scholar database and using only two keyword combinations. Further studies of this nature should include more keyword combinations, in more academic databases and choose a bigger timeframe.
Bibliography
Bearman, M., Ryan, J. & Ajjawi, R. Discourses of artificial intelligence in higher education: a critical literature review. High Educ 86, 369–385 (2022). https://doi.org/10.1007/s10734-022-00937-2
Ellingrud, Kweilin, et al. "Generative AI and the future of work in America." (2023).
Grassini, Simone. "Shaping the future of education: exploring the potential and consequences of AI and ChatGPT in educational settings." Education Sciences 13.7 (2023): 692.
Lim, Weng Marc, et al. "Generative AI and the future of education: Ragnarök or reformation? A paradoxical perspective from management educators." The International Journal of Management Education 21.2 (2023): 100790.
Miao, Fengchun, et al. AI and education: A guidance for policymakers. UNESCO Publishing, 2021.
Ng, Davy Tsz Kit, et al. "Conceptualizing AI literacy: An exploratory review." Computers and Education: Artificial Intelligence 2 (2021): 100041.
Ouyang, Fan, and Pengcheng Jiao. "Artificial intelligence in education: The three paradigms." Computers and Education: Artificial Intelligence 2 (2021): 100020.
Rampersad, Giselle. "Robot will take your job: Innovation for an era of artificial intelligence." Journal of Business Research 116 (2020): 68-74.
Salmon, Gilly. "May the fourth be with you: Creating education 4.0." Journal of Learning for Development 6.2 (2019): 95-115.
Webb, Michael, The Impact of Artificial Intelligence on the Labor Market (November 6, 2019). Available at SSRN: https://ssrn.com/abstract=3482150 or http://dx.doi.org/10.2139/ssrn.3482150
Williams, Peter. "Does competency-based education with blockchain signal a new mission for universities?." Journal of higher education policy and management 41.1 (2019): 104-117.
"107 Up-to-Date ChatGPT Statistics & User Numbers [Jan 2024]." N.p., n.d. Web. 8 Jan. 2024 https://nerdynav.com/chatgpt-statistics.
Acknowledgements
This study did not use ChatGPT or similar A.I. to write any content.
Thank you to the ResearchHub Community for the crowdfunded resources that allowed the author to commit the time to this study.
