Thirty years ago, economist Bengt-Åke Lundvall wrote that “the most fundamental resource in [a] modern economy is knowledge and, accordingly, the most important process is learning.” [1] Around the same time, Nelson Mandela commented that “eliminating the distinction between the information-rich and information-poor is… critical to eliminating economic and other inequalities between North and South.” [2]
Both men were alluding to an idea that took hold in the new Millennium, a new gap between rich and poor that quickly became known as the global Digital Divide. Health information technology has the potential to improve healthcare delivery and engagement. Studying these disparities in health information technology engagement will help understand and overcome challenges to healthcare utilization. The divide was defined by Lu as “great disparities in opportunity to access the Internet and the information and educational/business opportunities tied to this access … between developed and developing countries”. [3] At the turn of the century, 88% of all Internet users were from industrialised countries. Yet, those nations comprised only 15% of the world’s population. [4] In short, the world’s poor were being denied access to the greatest living repository of information that our species has ever known. And that denial would surely hamper their efforts to improve their economic status and equal health care accessibility. But since that observation, global internet use has increased dramatically. It grew from 20.6% of the world being online in 2007 to over 47% in 2016. [5] According to one tracker, today there are over 5 billion regular internet users, constituting nearly 65% of the global population. Of these, 4.9 billion are users of social media. [6]
The digital divide plays an important role in the distance gap between doctors and patients, also known as the broadband health gap. It is not uncommon to hear internet access being discussed as core social determinant of health. Certainly, the COVID-19 pandemic deepened this realization, with digital connectivity proving vital to both the continuity of medical care and the sustenance of public health surveillance and response systems, among other critical health roles.
The broad story of a global digital divide is not a simple one, of course, as there remains much regional and demographic heterogeneity. For instance, the overall gap in internet accessibility might be closing, but the regional gender gap remains wide: in 2016, there was a 23% internet use gap in Africa between women and men. [7] And while the gap between internet “haves” and “have nots” has been generally shortening, a Secondary Digital Divide was declared by some scholars, defined roughly as the gap between populations with and without high level information technology skills. [8] In other words, poor populations might be accessing the internet more, but they still lag in deep usage and in creating content for the internet, being largely dependent on rich populations for infrastructure, maintenance, content creation and curation. Related to this dependency is a lack of advanced use of the available technology, also considered to be a dimension of the Secondary Divide. For example, there are still dozens of countries where the majority of inhabitants might have access to the internet, but do not know how to attach a file to an email.
Some scholars have recognized a Tertiary Digital Divide, which alludes to the ability to use the internet in an effective and high-quality way. [9] The ability to identify appropriate and correct information in a soup of disinformation is one aspect of this phenomenon; it certainly is not limited to low-income populations and countries. This is particularly important given the acceleration of social media usage, as such platforms are complicit in the platforming and extreme amplification of falsehoods. The most recent digital transformation is the rise of Artificial Intelligence (AI), most immediately accessible to the general public as Large Language Models (LLMs), like chatGPT or Google’s Bard. These LLMs can produce grammatically perfect text, answer complex questions and perform high level text analysis, all while interacting with users in an easy conversational style. Their revolutionary nature cannot be understated. Despite only being available to the general public for a few months, LLMs are already transforming education, business, and general clerical work. Generative AI (of which LLMs are a type) are remaking multiple sectors, including artistic ones, by “pushing further into human realms” – which is a polite way of saying that LLMs can perform many high order human tasks faster and better than humans can.
The wealthiest nations and businesses are moving quickly to integrate this technology into their marketing and sales practices, their risk management activities, their research and development, and even their overall daily operations through the automation of such mundane tasks as to-do lists and summary reports. Lack of human expertise in language, organisation, and even some aspects of strategic thinking are no longer bars to high level institutional performance. The countdown is on until governments link large AI to critical functions currently managed by only the most trusted human overseers, such as the guidance of national economies, communication practices, and even militaries.
And the awesome potential of AI to transform healthcare cannot be understated, with machine learning transforming diagnostic processes, large language models promising to augment psychotherapists, and predictive algorithms poised to supercharge disease modelling and forecasting.
Given the undeniable competitive edge offered by these tools, institutions and governments that do not incorporate them will be at a distinct and widening disadvantage. Thus, it is time to declare a Quaternary Digital Divide: the gap between populations with functional access to generative AI and those without it.
This technology is, of course, not benign. It threatens to render countless categories of white-collar workers unemployed, and to usher in a new era of uncontrollable intellectual fraud and cybercrime. But like other disruptive innovations before it, such as history’s multiple industrial revolutions, its effects are likely unavoidable. And given its highly polarised distribution of access and uptake, it seems likely that AI’s immediate social effect will be to widen the gap rapidly and substantially between the global rich and the global poor –if its access is retained only by the former. As we scurry to understand AI’s impact on our own unprepared institutions locally, it behoves us then to also consider its grander context on the international development landscape. In the immediate term, LLMs can, for example, assist small NGOs in remaining competitive and functional despite diminishing human resources. And they can assist people with low literacy by summarising large, complicated documents into simple paragraphs more easily understood. But in the longer term, generative AI can be a transformative boon to emerging economies struggling with the Secondary and Tertiary Divides. An LLM diminishes the need for digital user expertise, given its nature as an automated personal assistant. For example, those populations mentioned earlier who are unable to attach a file to an email can theoretically simply instruct an AI to do so, using easy human language.
It is yet unclear whether generative AI is friend or foe. It’s probably both. It is evident, though, that the technology is scorchingly powerful and intractable. While many of us in wealthy nations resist its inculcation into our daily lives, we are well advised to accommodate its uptake and inclusion in less wealthy realms. Despite the very real and unsettling threats that AI presents, if well managed and well considered, AI just might be one of our best tools for bringing wealth, expertise, and parity to the world’s underserved populations.
REFERENCES
[1] Lundvall B. (1992). National Systems of Innovation. Towards a Theory of Innovation and Interactive Learning. Pinter, London and New York, pp. 342
[2] Wilson EJ. (2004). The Information Revolution and Developing Countries. MT Press, London.
[3] Lu M. (2001). Digital divide in developing countries. Journal of Global Information Technology Management (4:3), pp. 1-4.
[4] Pick J. & Azari R. (2008). Global Digital Divide: Influence of Socioeconomic, Governmental,and Accessibility Factors on Information Technology. Information Technology for Development, 14(2), 91-115
[5] Web Foundation. (2016). Digging into Data on the Gender Digital Divide. https://webfoundation.org/2016/10/digging-into-data-on-the-gender-digital-divide/
[6] Petrosyan A. (2023). Worldwide digital population 2023. https://www.statista.com/statistics/617136/digital-population-worldwide
[7] Nyamutswa C. (2018). Bridging the Gender Divide. International Telecommunication Union. https://www.itu.int/en/council/cwg-internet/Pages/display-oct2017.aspx?ListItemID=34
[8] Rubinstein-Avila E. (2016). “Diversification and Nuanced Inequities in Digital Media Use in the United States” in Handbook of Research on the Societal Impact of Digital Media, Guzzetti & Mellinee eds.,
Author
Prof. Raywat Deonandan, PhD
Prof. Raywat Deonandan is an Epidemiologist and Associate Professor with the Interdisciplinary School of Health Sciences at the University of Ottawa. He is also a researcher with the Centre for Health Law Policy & Ethics and the LIFE Research Institute, as well as a Senior Fellow with Massey College at the University of Toronto. Dr Deonandan holds a Research Chair in University Teaching, with a focus on the use of generative A.I. technologies for educational purposes in the health sciences.
Declaration of interests
I have read and understood the BMJ Group policy on declaration of interests and declare the following interests: None