Junhong Xiao
Emeritus Professor, Open University of Shantou

At the 29th World Conference of the International Council for Open and Distance Education (ICDE) last November, the president of an open university (OU) outlined the ambition of building a global digital university in his keynote speech. The promises of digital technology, including artificial intelligence (AI), for open and distance education are not new. A major issue of concern, though, is that people tend to be more interested in promise-making than promise-delivering. Promise-making is often romantically idealised while promise-delivering is fraught with concrete and contextual constraints. This is the case with AI for OUs.

The ambition for Asian OUs to deploy AI reminds me of the “racist” parkway bridges that Robert Moses (1888-1981) built in New York. Citing Robert A. Caro’s biography of Moses, Bloomberg reports that the bridges were deliberately designed to be extra-low “to keep buses from the city away from Jones Beach – buses presumably filled with the poor blacks and Puerto Ricans Moses despised” – so that convenient access to Long Island’s beach towns and parks on the Southern State Parkway became the privilege that only the “advantaged” who travelled by car instead of public transport system were entitled to. The same Bloomberg report describes Moses as “a tragic hero who built for the ages, but for a narrowly construed public.”

Drawing an analogy between AI systems and Moses’ “racist” parkway bridges may not be very appropriate in that the bridges were the result of purposeful discriminatory intention while AI systems are inherently meant for the public good. That said, if improperly or blindly adopted, AI can turn discriminatory and, in the case of OU education in Asia, may lead to more harm than good.

When it comes to AI for Asian OUs, the hard questions we ought to be asking are: How many Asian OUs can boast AI applications? Even if Asian OUs can afford to deploy AI systems for the purpose of learning and teaching, how many of their learners can actually afford or access AI applications in their daily contexts?

Promise-making is often romantically idealised while promise-delivering is fraught with concrete and contextual constraints. This is the case with AI for OUs.

The fundamental mission of OUs is to contribute, in the first place, to higher education for all, in particular for those who are ethnically, geographically, economically, socially, physically, or even sexually disadvantaged. Although OUs, especially in developed countries or metropolitan areas, tend to cater for an increasing number of “advantaged” students today, they cannot claim to be fulfilling their mission unless the “disadvantaged” cohorts, no matter how insignificant in size, do not remain disadvantaged or are not even further marginalized as a result of the institution’s digital transformation.

The demand for affordable and hence equitable access to higher education in Asia has always been enormous, turning Asia into the home of the world’s largest OUs although this model of education was not Asia-native. And lest we forget, based on the threshold of USD$3.20 per person per day, “the Asia-Pacific region is still home to half of the world’s poor people”, according to the United Nations’ Economic and Social Survey of Asia and the Pacific 2022. The Key Indicators for Asia and the Pacific 2023 released by the Asian Development Bank shows that 3.9% of the Asia-Pacific region’s population lived in extreme poverty line, that is, on less than USD$2.15 per person per day in 2022.

if improperly or blindly adopted, AI can turn discriminatory and, in the case of OU education in Asia, may lead to more harm than good.

Given that financial difficulties are a key factor leading to poor access to (higher) education, no efforts have been spared in the search for cost-saving or cost-effective models of education whether at the level of policy-making or in research and practice. Of all the options, technology-based provision has always been regarded as the most promising silver lining, if not the panacea for equitable access to (higher) education. In addition to its promises to enhance educational quality, technology is often assumed to be able to reduce cost, hence arguably making education accessible to more people. The latest hype surrounds the most cuttingedge technology of the day, that is, AI.

Putting aside the lack of large-scale rigorous empirical evidence of AI’s effectiveness in improving educational quality, or to be more specific, the quality of OU education, there is solid evidence that the costeffectiveness rhetoric of AI goes against the fact. The research and development (R&D) of AI is expensive and so is the usage stage. In other words, the consumers, i.e., educational institutions and students, have to shoulder the cost incurred not only when they are using AI systems but also during the R&D stages. Developers will no doubt transfer the R&D cost to the final products. Moreover, an AI system is not a one-off investment; a continuing flow of money is needed to sustain its operation, not to mention that this operation also depends on the availability of relevant adequate infrastructure as well as qualified personnel, among other things.

Therefore, only educational institutions in wellresourced countries can benefit from AI systems (if they are indeed useful), further exacerbating inequities in education in poorly-resourced countries. To quote Professor Maria Mercedes T. Rodrigo of Ateneo de Manila University, whom I cited in the previous issue of inspired, those “who stand to benefit the most from AI-powered education” are usually not able to reap its rewards, as is evidenced by the situation in the Philippine context. This begs the question of whether AI can be an empowering tool for OUs in Asia to better fulfil their mission.

As is obvious from the above analysis, on the whole, Asia is still relatively less developed, especially when compared with Western Europe and North America. Therefore, unlike OUs in developed countries, many OUs in Asia may not be able to obtain sufficient financial support for investment in AI from their governments, including constructing the necessary infrastructure to deploy AI-based educational applications at scale. The situation may be contrasted with the OU in the United Kingdom (OUUK), which recently received a grant of £5.8m from the Office for Students to build extended reality studios with the aims of expanding its teaching and learning with augmented and virtual reality, and creating authentic contexts for skills development. The OUUK is also experimenting with the use of augmented reality in language learning and an immersive virtual environment of a modern courtroom to facilitate student learning. Similar instances of the deployment of cutting-edge technologies in OUs in developed countries are too numerous to list here.

In light of the above arguments and evidence, we have no reason to be optimistic about the answers to the hard questions we asked earlier. It is not realistic to expect that many Asian OUs can boast such AI applications which are commonplace to their counterparts in developed countries. Even if they can, it is not realistic either to expect that the majority of their learners, especially those “who stand to benefit the most from AI-powered education” in the eyes of Professor Rodrigo, can be the beneficiaries of AI in education due to the obvious cost implications involved.

It is worth noting that the equity discourse is embodied in the grand narratives of official documents, for example, UNESCO’s AI and Education: Guidance for Policy-Makers, the OECD’s Shaping Digital Education: Enabling Factors for Quality, Equity and Efficiency, and the US’ Office of Educational Technology’s Artificial Intelligence and the Future of Teaching and Learning: Insights and Recommendations. In contrast, the ambition of promoting equity is rarely mentioned by major companies specializing in AI in education, as pointed out by Paulo Blikstein et al. in their research article titled “Discourses of Artificial Intelligence in Education through the Lens of Semiotic Analytics” (2022). This lack of equity-related discourse is by no means due to the negligence of these commercial players because they know very well that the role of AI in promoting equity in education is limited.

Unless adequately informed by sound analysis and solid evidence, any attempt to jump on the AI bandwagon will likely privilege ‘a narrowly construed public’

Therefore, we should be cautious when making decisions on the adoption of expensive cutting-edge technologies such as AI for the purpose of learning and teaching. Unless adequately informed by sound analysis and solid evidence, any attempt to jump on the AI bandwagon will likely privilege “a narrowly construed public” as in the case of Robert Moses’ low parkway bridges, that is, the “already advantaged” for whom AI in education is the icing on the cake, rather than the disadvantaged who stand to benefit the most from AIbased education, hence leading to unintended harmful consequences despite the best intentions.

Remember that any decision-making concerning AI should be contextualized and made out of real necessity. We should guard against the desire to follow suit blindly or out of fear of missing out. Otherwise, we would be no different from Robert Moses.

Prof Junhong Xiao may be contacted at