Tagging AI as Divisive Misses the Bigger Picture

Of the many issues raised by the advance of AI in the workplace, one of the most contentious is that it could deepen the divides that separate the world’s haves from the have-nots. A new report from the IMF warns of these deepening inequalities.

AI-driven disparities are a cause for concern. However, the picture may be more nuanced than the IMF's, and we need to be thoughtful about the future dynamics of job markets and AI’s social impacts.

Diverging employment paths

In its analysis, the IMF acknowledges that the net effect of AI’s impact is difficult to foresee and will ripple through economies in complex ways. However, the organization is less equivocal on the inequalities AI could create. “In most scenarios, AI will likely worsen inequality, a troubling trend that policymakers must proactively address to prevent the technology from further stoking social tensions,” says the IMF.

On the labor front, the report finds that almost 40% of global employment is exposed to AI. However, the nature of this exposure – and hence the potential for widening social divides – varies across the employment universe.

The IMF says that in advanced economies with more high-skilled employment opportunities, roughly half the vulnerable jobs may benefit from AI integration, enhancing productivity. If this happens, these high-income workers will benefit from disproportionate earnings increases. Moreover, the productivity gains will boost employers’ capital returns, which could also favor high-income employees.

It is a different story for the other half of workers exposed to the vagaries of AI. Here, the technology could lower labor demand by taking over critical tasks currently performed by humans. In some cases, jobs may disappear. The net result for these employees is likely to be lower incomes and fewer job openings.

This polarization also has an age component. According to the IMF, research shows that AI can help less experienced workers enhance their productivity more quickly. Younger workers may find it easier to harness the benefits of AI than their older counterparts.

The outlook outside of advanced economies is different, argues the IMF. In emerging markets and low-income countries, some 40% and 26% of jobs, respectively, are vulnerable to AI’s influence. The implication here is that employment markets in these countries face comparatively fewer immediate disruptions from AI.

National inequities

The IMF analysis suggests that the potential for AI-induced inequalities could extend to countries as well.

Since more jobs in advanced economies – some 60% estimates the IMF – than in emerging and low-income countries are affected by AI, the commensurate risks of widening social divides are not as great in latter nations.?

However, unlike advanced economies, emerging and low-income countries lack the infrastructure or skilled workforces to capture AI's benefits fully, maintains the IMF. This disparity could open rifts between nations at different stages in their economic development.

Counter arguments: jobs

While the IMF’s analysis is thought-provoking and merits attention, a contrarian view is that AI can also act as a wealth equalizer.

Consider the argument that higher-paid, white-collar workers could benefit more than lower-paid, blue-collar workers by capturing AI’s productivity-boosting premium. Another scenario is that in an AI-enabled world, many white-collar jobs, for example, those involving computer work, are more replaceable than blue-collar jobs like plumbing and food service. In this sense, AI levels the economic playing field.

It is also notable that AI is more widely available than other revolutionary technologies like electricity and computers when they emerged. Initially, these technologies were expensive and available primarily to wealthier users. AI, on the other hand, is accessible to anyone who has a smartphone, probably about 75% of the world’s adult population. Indeed, AI chatbots are accessible to anyone with a basic cellphone via text messages.

Another area in which AI promotes equal opportunities is the cost and availability of education. Digital technologies and platforms like MIT’s MOOCs (massive open online courses) are lowering the cost of advanced education and making it available to vast audiences. AI supports this revolution in various ways, such as facilitating the increasing use of personal tutors, refining curricula, augmenting course materials, and providing automatic translation services.

AI also has an equalizing role in the corporate world. As I argue in my book The New (Ab) Normal: Reshaping Business and Supply Chain Strategy Beyond Covid-19 (MIT CTL Media, 2020), the rise of relatively inexpensive, cloud-based business services enables under-funded small and medium-sized enterprises (SME) across the globe to adopt cutting-edge technology. AI’s cost model helps cloud service providers serve small business markets and SMEs to become adopters.

Counter arguments: Countries

AI’s impact on national economic differences is similarly nuanced.

The IMF analysis notes that job markets in wealthier nations are likely to be impacted more severely by AI than those in poorer nations. This outcome implies that less affluent countries largely decoupled from the knowledge economy will, to some extent, be shielded from the turmoil in knowledge work.

Moreover, poorer countries with economies heavily reliant on exporting natural resources and tourism stand to benefit from the productivity-fueled growth in advanced economies made possible by AI since many of their customers reside in wealthier countries.

Another factor is AI’s ability to even out the competitive landscape globally. As described above, the growth in AI applications helps SMEs worldwide to adopt innovative technologies that improve their competitiveness and make it easier for them to compete internationally. An example is the adoption by poorer nations' SMEs of the same cloud-based services used by companies in advanced economies, assuming local and national governments do not restrict access to these services. Automated language translation services will give suppliers in poorer countries more access to customers in more affluent parts of the world.

One way in which AI might widen competitive gaps between countries is by reversing the offshoring juggernaut that led to jobs being exported from wealthier to poorer nations. For example, using call centers and business process outsourcing expertise in India makes little sense in a world where AI handles this work. Why hire inexpensive Indian or Eastern European programmers when AI-augmented programmers in Western countries are more productive? Also, in some countries, notably China, there are fewer ethical and civic obstacles to the spread of automation.

Consensus on government action

Regardless of how AI’s impact on inequality pans out, one thing is for sure: governments need to be mindful of the risks and start planning for an AI-driven workplace. I argue this in my book The Magic Conveyor Belt: Supply Chains, AI, and the Future of Work (MIT CTL Media, 2023).

The IMF’s analysis agrees: “It is crucial for countries to establish comprehensive social safety nets and offer retraining programs for vulnerable workers. In doing so, we can make the AI transition more inclusive, protecting livelihoods and curbing inequality.”