A thought-provoking question has been raised about the potential impact of artificial intelligence (AI) on society. What if AI solves all of science's hardest problems, but in doing so, exacerbates existing social and economic inequalities? This scenario was explored in a recent article by Steven Boykey Sidley, which posits that even if AI brings about unprecedented technological advancements, unequal access and ownership may lead to a widening of the gap between the haves and have-nots.
According to Sidley, the issue lies not with the technology itself, but with the way it is distributed and accessed. He cites the work of economists Diego Comin and Bart Hobijn, who tracked the adoption of 15 technologies across 166 countries over two centuries. Their research revealed that while the time it takes for technologies to reach poor countries has decreased, their penetration into these societies remains limited. This has contributed to what Comin and Martí Mestieri term the "Great Income Divergence" since 1820.
The problem of unequal access to AI is further complicated by the issue of ownership. If a small group of companies and investors owns the models, data centers, robot fleets, and intellectual property that produce AI, then the benefits of this technology may be concentrated among the wealthy. This could lead to a situation where AI simultaneously alleviates physical drudgery and increases economic inequality. Furthermore, some assets, such as land, housing, and education, will remain scarce, and their value may rise if AI makes other goods and services cheaper.
Another challenge posed by AI is its potential to create self-amplifying inequality. Suppose an AI-designed longevity treatment is developed, but only accessible to the wealthy at first. Those who can afford it will gain a significant advantage over others, which may be difficult to reverse even if the technology becomes more widely available later. This compounding effect could apply to various areas, including education, cognitive enhancement, and wealth management.
The issue of unequal access to AI has significant implications for the future of society. If AI is not distributed fairly, it may favor those with the capital to own the models, the compute to run them, and the skills to use them. This could lead to a widening of the gap between the AI-haves and AI-have-nots. However, it is worth noting that AI may still bring about significant benefits, even if access is limited. For example, a woman in Soweto with a cheap Android phone and a decent model may have better medical triage than a Harley Street patient had a decade ago.
While AI may not create a binary divide between the haves and have-nots, the benefits of this technology are likely to be distributed unevenly. Those with means will be able to access more advanced AI tools and applications, while those without will be limited to more basic versions. This is consistent with the Comin-Mestieri pattern of convergent arrival and divergent penetration.
Ultimately, the impact of AI on society will depend on how it is developed, distributed, and accessed. If steps are not taken to address the issue of unequal access and ownership, AI may exacerbate existing social and economic disparities.
Key points
- AI may solve science's hardest problems, but unequal access and ownership may deepen existing social and economic inequalities.
- The issue of unequal access to AI is complicated by the concentration of ownership among a small group of companies and investors.
- AI may create self-amplifying inequality, where those who have access to this technology gain a significant advantage over others.
