ING AI Report
AI may equalise opportunity, but inequality could worsen first
AI lifts productivity and access for smaller teams
Gains concentrate early, risking wider inequality gaps
Stock video by João Adão
AI as the Great Equaliser: Potential and Pitfalls
Artificial intelligence holds the potential to become the Great Equaliser of the twenty-first century, according to a report from ING. The analysis examines both short-term and long-term economic impacts of AI, focusing on productivity, capability, and inequality.
Productivity Gains and Their Limits
The report notes that AI’s impact on macro-level productivity has been limited so far, despite notable micro-level improvements. Tasks such as coding, translation, writing, and data analysis can be completed much faster, with survey data suggesting improvement gains of around 50% in many areas. As agentic AI usage grows, productivity improvements are expected to increase further as entire functions become automated.
AI also enables people to work smarter by reducing time spent on routine tasks, allowing more focus on higher-value work. Additionally, AI can give time back to individuals who might otherwise need to pay with time, potentially helping to reduce inequality.
However, productivity gains are likely to be asymmetric. Among AI adopters, a cohort of superusers may dramatically scale their output, while others may misuse the technology, creating accuracy and reliability issues. On average, AI is expected to allow users to produce higher quantity and quality output, bridging the gap between resource-rich large organisations and smaller teams.
Higher productivity alone does not guarantee greater equality. Historically, productivity improvements have accrued to capital owners rather than labour. If AI follows a similar pattern, workers may not capture the gains.
Expanding Capability and Opportunity
The second transformative impact of AI is enabling people to do things previously out of reach. AI supercharges the democratisation of information and expertise, building on the internet’s role in bridging knowledge gaps. It helps offset skill and knowledge deficits, allowing workers to compensate for areas where they are less proficient. For example, a weak writer can use AI to communicate with professional clarity, and those without technical skills can use natural language prompts to complete complex tasks.
AI also lowers barriers to entrepreneurship. Historically, funding has been a defining bottleneck, but with AI, one-person or resource-light companies become increasingly viable. The barriers to entrepreneurship are likely to shift from capital-constrained to creativity-constrained, potentially reducing inequality as entrepreneurship remains a key avenue for socioeconomic mobility.
The report suggests that AI’s role in expanding capability and opportunity may play a larger role in improving equality globally than productivity gains alone.
Inequality May Worsen Before It Improves
Despite the positive long-term outlook, the report argues that inequality is likely to worsen before it gets better. Four key arguments support this view.
First, gains from the AI race are currently accruing to select companies and economies. A PWC study concluded that 75% of AI gains are captured by just 20% of companies. This has led to clear winners and losers, with increasing divergence in growth across countries. For example, in China, hi-tech sectors attract disproportionate capital and resources, resulting in a K-shaped economy where a handful of tech-driven sectors expand explosively while the traditional economy stagnates. Unlike previous cycles, tech and AI supply chains are relatively shorter, so fewer people benefit from the tech boom compared to earlier real estate and infrastructure investment supercycles.
Second, governments face difficult questions about intervening in distributing AI gains. For instance, Taiwan paid out NT$10,000 per person in 2025 and will do so again in 2027 amid the AI boom. A national dividend proposal in South Korea in May suggested redistributing revenue from the semiconductor boom to citizens, which shook markets.
Third, workers may become more productive but lose bargaining power. Companies may be more inclined to spend on tokens than talent, potentially reducing workers’ bargaining power. Wage growth has slowed in major economies such as the US, China, and the EU in recent years, though it is too early to conclude causality.
Fourth, the AI transition could hamstring a generation of workers. Job displacement is a common concern, with AI cited as a leading reason for US job lay-offs in recent months. Less discussed is the decrease in entry-level hiring, as AI easily replaces much entry-level work. This could have significant implications for lifetime earnings and future consumption. A World Bank working paper argues that AI adoption has a greater negative impact on labour demand for entry-level and lower-educated workers. Youth unemployment rates are already high, and AI’s introduction comes at an unfortunate time. While it is too early to draw firm conclusions, youth unemployment has edged higher in Asia and North America, while falling in the Eurozone.
The report also notes that AI-linked robotics are gradually coming to market, potentially replacing dangerous, boring, or undesirable jobs, but also highly paid or desirable ones. Every major technological shift has created new occupations and rendered others obsolete, and AI is likely no different. Government handling of worker retaining and employee protection will determine how much this amplifies inequality.
Geographical Disadvantages Persist
AI will do little to address entrenched geographical disadvantages. Even if AI proves to be a drastic paradigm shift, it will not help those without access. The International Telecommunication Union estimates that around a quarter of the global population still lacks reliable internet access. Internet penetration is over 90% in North America and Europe, but just 35% in Sub-Saharan Africa. A good proportion of people in frontier economies will likely also be without access to AI and its benefits, before considering token costs. This will likely worsen inequality before things get better.
Conclusion: Potential for Equalisation, but a Difficult Road
The report concludes that AI does have the potential to become the Great Equaliser in time. AI is already increasing productivity, helping workers overcome skill and knowledge gaps, and is likely to lower barriers to entrepreneurship and improve access to opportunities globally. When the dust settles, the playing field may be more even than before the AI era.
However, historical evidence suggests that inequality will likely get worse before it gets better. Early beneficiaries of technological shifts tend to be concentrated among capital owners, and broader social gains take years or decades to emerge. There will also likely be demographics left behind, as with previous waves of technology.
Whether this vision materialises will depend on how AI is regulated, which development model wins the race (open versus closedistributed. AI will not make everyone equally wealthy, but if it succeeds in making expertise, entrepreneurship, and opportunity more accessible, it may become one of the most important forces for levelling playing fields in the twenty-first century
Recommended posts
Free Data: Insights – World
Instant access. No credit card needed.
