Global brokerage firm Bernstein has challenged the conventional wisdom about AI's economic impact, arguing that developed nations will bear the brunt of artificial intelligence disruption. According to Bernstein's research note, the most common narrative suggests that knowledge industries will be hardest hit, entry-level roles will disappear first, and low-cost labor economies like India, Indonesia, and Mexico face the greatest threat. However, the firm argues this perspective reveals a striking disconnect with actual economic data. The firm's analysis, written by Venugopal Garre and Nikhil Arela, challenges the common view that AI will hurt middle-income countries like India the most, arguing this perspective is mistaken.
The research firm's analysis reveals significant sectoral exposure differences between middle-income and high-income economies. As reported by Bernstein, information and knowledge sectors account for only 1-4% of employment in most middle-income economies versus 6-10% in high-income ones. The GDP share gap is even wider, with services contributing about 50-60% of output in middle-income countries (median 55%) compared to more than 70% in advanced economies. This structural difference means AI's impact will be more concentrated in developed nations where these sectors represent a larger portion of the economy.
Country-specific wage differentials provide a compelling economic rationale for AI adoption patterns. According to Bernstein's analysis, a typical software engineer earns around ₹8.3 lakh annually in India compared to roughly ₹1.07 crore in the US and ₹6.4 lakh in the UK. This wage gap extends across white-collar professions including accountants, paralegals, and data analysts where AI growth has been most conducive. As Bernstein explains, replacing a ₹8.3 lakh worker with AI yields only marginal savings, while replacing a ₹1.07 crore worker is transformative for most organizations, creating strong economic incentives for AI adoption in developed nations. The report notes that if firms adopt human-in-the-loop systems that retain a limited pool of decision-makers, the question becomes which humans make the most economic sense to keep.
Based on current service-sector employment and existing income inequality, Bernstein estimates the economic impact will vary significantly between income levels. The research firm projects an average 10% hit to national income in middle-income economies compared to about 22% in high-income economies. This differential reflects the concentration of AI-susceptible sectors and higher-paid employment in developed nations. The firm notes that AI may achieve a narrower gap between rich and middle economies through two dimensions: reducing income gaps between high- and middle-income countries, and flattening income distributions beyond the top 1% within economies. The burden will fall disproportionately on the next-richest 9-10% income cohort after the top ~1%, making AI a leveller along these dimensions.
Despite Bernstein's counter-intuitive analysis, markets have reacted differently, with stock markets in India and Indonesia suffering the most from AI-related concerns. As the report explains, markets are short-sighted, focusing on the next few quarters rather than decades. India has high exposure to IT services in its benchmark indices (around 10%), while emerging markets lack big pure-play AI companies. However, Bernstein suggests that near-term disruption to India's IT services may have been over-discounted, noting that IT vendors will still be needed to help companies adopt AI, which could support margins. The firm expects AI to become a tightly regulated industry with potential minimum employment mandates, restrictions on AI usage, Universal Basic Income, or taxes on AI adoption.
