Artificial intelligence (AI) is more than a buzzword these days. It powers facial recognition on smartphones and computers, translation from foreign languages, systems that filter spam in emails and the identification of toxic content on social networks, as well as detecting cancerous tumors. These examples, along with numerous other existing and emerging applications of AI, help make everyday life easier for people, especially in developed countries.
As of October 2021, 44 countries have reported having their own national AI strategic plans, demonstrating their willingness to move forward in the AI race. These include emerging economies such as China and India, which are leading the way in drawing up national AI plans in the developing world.
Oxford Insights, a consulting firm that advises organizations and governments on issues related to digital transformation, ranked 160 countries around the world prepared for the use of AI in public services. The US ranked first in the 2021 AI Government Readiness Index, followed by Singapore and the UK.
Notably, the regions ranked lowest on the index include much of the developing world, such as sub-Saharan Africa, the Caribbean and Latin America, as well as Central and South Asian countries.
The developed world has an unavoidable advantage in rapid progress in the AI revolution. With greater economic capacity, richer countries are naturally well positioned to make the high investments needed in research and development to create modern AI models.
By comparison, developing countries often have more pressing priorities, such as education, sanitation, healthcare and food, that override any significant investment in digital transformation. In this climate, AI can widen the digital divide that already exists between developed and developing countries.
The hidden costs of modern AI
AI is traditionally defined as “the science and engineering of producing intelligent machines”. To solve problems and accomplish tasks, AI models often look at past information and learn rules for making predictions based on the unique patterns present in the data.
AI is a broad term, covering two main areas: machine learning and deep learning. While machine learning tends to be suitable when it involves learning from smaller, well-organized datasets, deep learning algorithms are better suited for complex, real-world problems such as predicting respiratory disease at from chest X-ray images.
Many modern AI applications, from Google’s translation function to robot-assisted surgical procedures, employ deep neural networks. They are a special kind of deep learning model loosely based on the architecture of the human brain.
A crucial aspect is that neural networks are data-hungry, often requiring millions of examples to learn how to do a new task well. This means they require complex data storage infrastructure and modern computing hardware compared to simpler machine learning models. Such large-scale computing infrastructure is often unaffordable for developing countries.
In addition to the high price, another problem that disproportionately affects developing countries is the high cost this type of AI has on the environment. For example, a contemporary neural network costs over $150,000 to train and will produce around 650 kg of carbon emissions during training (something comparable to a trans-American flight). Training a more advanced model can lead to approximately five times the total carbon emissions generated by an average car over its lifetime.
Developed countries have historically been the main contributors to the increase in carbon emissions, but the burden of these emissions unfortunately falls most heavily on developing countries. The Southern Hemisphere often experiences disproportionate environmental crises such as extreme weather, droughts, floods and pollution, in part because of its limited ability to invest in climate action.
Developing countries also benefit less from advances in AI and all the good they can provide, including building resilience against natural disasters.
Using AI for good
While the developed world makes rapid technological progress, the developing world seems underrepresented in the AI revolution. And in addition to uneven growth, the developing world is likely to bear the environmental consequences created by modern AI models, most of which are employed by the developed world.
But it’s not all bad news. According to a 2020 study, AI can help achieve 79% of sustainable development goals. For example, AI could be used to measure and predict the presence of contamination in water supplies, thus improving water quality monitoring processes. This, in turn, could increase access to clean water in developing countries.
The benefits of AI in the Southern Hemisphere could be vast, from improving sanitation, helping education, providing better medical care. These incremental changes could have significant flow effects. For example, better sanitation and better health services in developing countries could help prevent disease outbreaks.
But if we are to realize the true value of “good AI”, equal participation in the development and use of technology is essential. This means that the developed world needs to provide greater financial and technological support to developing countries in the AI revolution. This support cannot be short-term, but it will create significant and lasting benefits for everyone.
This article is authored by Joyjit Chatterjee, Data Scientist (Knowledge Transfer Partnership Associate), University of Hull, and Nina Dethlefs, Senior Lecturer in Computer Science, University of Hull, reproduced from “The Conversation” under license from Creative Commons.
The authors do not work, consult with, hold equity interests in, or receive funds from any company or organization that would benefit from this article and report that they have no relevant affiliations beyond their academic commitment.