Image made on Canva Pro by Ameya Nagarajan for Global Voices.
By Aaron Spitler
Much ink has been spilled about how artificial intelligence (AI) might transform the global economy. A recent report from UN Trade and Development (UNCTAD) projected that the AI market will be valued at USD 4.8 trillion by 2033. For proponents of AI, this forecast shows how the technology will be a boon to all countries. However, the organization reached a different conclusion, namely that the sector’s growth would primarily benefit the global north. Numerous barriers, from inadequate technical infrastructure to insufficient educational opportunities, have prevented countries in the global majority from reaping the rewards of the AI revolution. Looking beyond the hype, the reality may be that this divide only worsens. It is worth examining how AI may split the world into haves and have-nots.
At the micro-level, the adoption of AI across industries spells trouble for workers in lower-income countries. More specifically, the automation of tasks classified as repetitive or mundane could lead to the elimination of jobs at scale, thrusting the livelihoods of thousands throughout the global majority into jeopardy. The challenges posed by AI to the global majority do not end with economic disruptions. Zooming out, many AI tools being adopted by the public were created in the global north, reflecting the needs and priorities of a select few. Troublingly, these solutions have also been trained on data rife with biases about the global majority, digitizing stereotypes and tropes that have lingered for centuries. Ultimately, without action, the AI boom may prove to be a bust for less-resourced nations.
Fomenting economic uncertainty
Concerns about how AI will shake up the labor force are widespread. Yet its impacts may be disproportionately felt in the global majority. Foreign Policy highlighted how companies in these countries, fixated on cost savings and efficiency gains, are assessing how AI can replace labor provided by humans. Jobs that do not require significant expertise are often the first on the chopping block. As a result, workers in the global majority who lack in-demand skills may lose the roles they depended on. Compared to their peers in the global north, opportunities for these individuals to boost their capacities are rarer. Facing scant options for upskilling, as well as an increasingly competitive job market, workers in the global majority may soon find themselves in a precarious position.
The strain this economic upheaval may place on the social safety nets of these countries also deserves attention. A blog from the Center for Global Development (CGD) touched on this point, noting that many nations do not have robust or reliable safeguards in place to prevent citizens from enduring financial hardship. When spikes in unemployment shock under-funded and under-developed social protection systems, their ability to support citizens out of work is stretched to the limit. As a result, the likelihood that many slip through the cracks to face poverty increases, creating a problem that affects large swaths of the general population. Although some may downplay AI-related job loss as inevitable, the second-order effects it may have on the resilience of a society should not be overlooked.
Digitizing cultural hierarchies
Issues with AI tools are not only found in how they are deployed; many can also be traced back to how they were developed. Project Syndicate made this case, arguing that the AI-powered technologies that have become mainstream were designed for consumers based in the global north. Importing them to countries they were not built for, without accounting for factors like cultural context, diminishes their efficacy. Therefore, the chance that these “solutions” end up causing more problems for users in the global majority may be higher than acceptable. While some countries have begun crafting regulations designed to minimize AI-related harms, many do not yet have the policy infrastructure to take on this multifaceted challenge. In these situations, the damage may already be done.
AI-enhanced products created in the global north that later spread across the global majority are flawed in more ways than one. As Wired identified, they are also encoded with the values of those from these regions, disseminating a worldview which may not align with the experiences of those from the global majority. These tools, trained on data that presents a narrow representation of life in the global majority, are sold to consumers based in these countries as an asset. In actuality, they are liabilities, promulgating biases and falsehoods about certain communities in the outputs they generate. With few pathways for recourse, critics of Western-centric AI tools often struggle to present alternatives, knowing that their resources pale in comparison to firms located in Europe or North America.
Imagining equitable solutions
Assessing the evidence, it is reasonable to conclude that AI has the potential to exacerbate inequalities affecting the global majority. Even so, this outcome is not set in stone. To address how AI may upturn the lives of individuals, particularly in disrupting their livelihoods, policymakers from these regions must take action to help their constituents. Launching skill-building initiatives and strengthening social programs, for instance, may go a long way towards helping citizens climb out of economic vulnerability. Furthermore, establishing spaces where leaders across the global south can discuss rules and standards for AI should be prioritized, as these initiatives would enable officials in government to deepen their expertise. Importantly, insights from these fora could be integrated into regulations that center on the needs of individuals, ensuring that policies are responsive to their most pressing concerns.
Building AI tools that are precisely calibrated to the global majority would also make a meaningful difference. To reach this goal, nurturing the creation of a digital ecosystem in these regions, with support for grassroots innovators, is a must. By working together, countries in the global majority may gain the capacity to challenge the dominance of the global north in this field, whose imperfect (if not harmful) tools have flooded their markets. This collaboration may result in a proliferation of tools that are better tailored to contexts across the global majority, all while shrinking the pool of products that replicate stereotypes about these communities. Measures like these may help narrow the AI gap that exists in our international community, giving rise to a world where this technology truly works for all.
Aaron Spitler is a researcher whose interests lie at the intersection of emerging technologies and human rights. He has worked at a number of organizations in the technology policy space, including the International Telecommunication Union, the Internet Society, and Harvard’s Berkman Klein Center. He is passionate about ensuring technology can be used as a force for good.