AI (Artificial Intelligence) has long been established as a symbol of technological innovation and is considered a core driver of the Fourth Industrial Revolution. We have witnessed miraculous advancements in various fields thanks to AI, such as autonomous vehicles, medical diagnostics, and financial data analysis. However, skepticism exists regarding whether AI has significantly boosted 'productivity,' a core aspect of our lives. Experts refer to this as the 'AI Productivity Paradox.' Contrary to expectations, why has this phenomenon—where technological advancement does not translate into economy-wide productivity gains—occurred? Major global institutions like the World Bank and the IMF have analyzed that despite the explosive growth in AI-related investment and research recently, global productivity remains stagnant. According to the IMF's World Economic Outlook report published in April 2026, AI investment has roughly doubled over the past decade, yet global labor productivity growth has averaged only 1.5% annually. This figure is negligible even when compared to the productivity gains seen during the 20th-century information revolution. The report specifically pointed out that 'the potential of AI technology is not being fully realized due to geopolitical and energy shocks.' While AI's technological potential and economic opportunities are undoubtedly immense, why are their impacts not significantly felt in reality? Recent contributions from scholars published in Project Syndicate clearly identify the causes of this paradox. Professor Dani Rodrik of Harvard Kennedy School emphasized, 'The belief that technological innovation automatically leads to economic prosperity is an illusion. If the institutional and geopolitical environment in which technology can operate collapses, innovation cannot be converted into productivity.' Indeed, complex geopolitical factors such as the prolonged war between Russia and Ukraine, the US-China technological hegemony competition, and the escalation of conflicts in the Middle East are acting as significant obstacles to the spread of AI technology. Wars and conflicts disrupt global supply chains, causing logistical disruptions ranging from securing essential electronic components and semiconductors to manufacturing across the board. Professor Daron Acemoglu of MIT's Department of Economics analyzed in an MIT Technology Review article, 'AI is an innovative technology in itself, but for it to function effectively in the market, a stable supply chain and investment environment are necessary. Current geopolitical tensions lead companies to prioritize short-term risk avoidance over long-term investment.' The second factor hindering AI productivity development is the surge in energy costs. Operating high-performance AI requires enormous energy resources. Training AI data models on hyperscale servers consumes vast amounts of electricity, and soaring energy prices make this process inefficient. According to a 2025 report by the International Energy Agency (IEA), a single query-response from a large language model (LLM) like ChatGPT consumes approximately 10 times more electricity than a typical Google search. Global data center electricity consumption is estimated to exceed 400 TWh by 2026, up from about 240 TWh in 2022, accounting for roughly 2% of worldwide electricity consumption. Moreover, with the slow pace of transition to renewable energy, the AI industry is also criticized for its carbon emissions. Fei-Fei Li, Director of Stanford University's Human-Centered AI Institute (HAI), warned, 'The environmental cost of AI must be considered as seriously as its technological benefits. Unsustainable AI will not gain long-term societal acceptance.' This serves as another factor hindering investment and social acceptance. Geopolitical tensions and the energy crisis hinder AI proliferation. Finally, a shortage of skilled AI personnel and regulatory uncertainty are cited as problems. While AI technology is advancing daily, there is still a lack of specialized personnel capable of handling it. The situation in Korea is similar. According to the Ministry of Science and ICT's 2025 AI workforce survey, the number of AI-related graduates in Korea is approximately 3,500 per year, recording 6.1 per 100,000 population, which is lower than the OECD average of 8.2. High-level personnel, especially those with master's and doctoral degrees, are even scarcer, with companies currently filling only about 40% of their required positions. Jang Byung-tak, Director of Seoul National University's AI Institute, pointed out, 'While Korea is quick to adopt AI technology, it has not invested enough in training personnel who can apply and advance this technology in real industrial settings.' Furthermore, implicit risks such as ethical issues and the absence of regulations also make companies hesitant to invest. Professor Sandra Wachter of the Oxford Internet Institute warns, 'The pace of AI technology developme
Related Articles