Experts in artificial intelligence foresee significant economic transformations driven by advancements in AI technology, with some envisioning rapid growth and prosperity in the coming years. Sir Demis Hassabis, founder of DeepMind, has described the emerging period as “an era of radical abundance,” while technology investor Marc Andreessen predicts “absolutely stratospheric” productivity gains that could render concerns about inequality obsolete. Additionally, Anthropic’s economics team forecasts a median global GDP growth rate of around 8 percent annually by 2030.
Despite this optimism, analysts caution that the surge in AI capabilities may not necessarily translate into proportional increases in measured GDP or improvements in human welfare. One key concern is that as economies develop, a growing portion of activity becomes zero-sum, involving competition where gains for some come at the expense of others. For example, AI-enabled cyberattacks are expected to become more sophisticated, leading to escalating cyber defense costs that do not directly benefit consumers. Similarly, while AI may automate certain legal tasks, it could also drive a substantial rise in litigation. This phenomenon—sometimes referred to as “agentic flooding”—has already overwhelmed the British legal system with AI-generated complaints and injunctions.
On the other hand, AI promises breakthroughs that could greatly enhance human welfare at little cost. Advances in drug discovery accelerated by AI techniques might eventually yield medications capable of ensuring long and disease-free lives. Initially, these innovations could boost GDP figures through increased research, production, and pharmaceutical profits. However, once patents expire and manufacturing becomes highly automated, the economic contribution and associated earnings could sharply decline. At the theoretical extreme, goods made infinitely abundant by AI would have an almost negligible measurable economic value, despite substantial welfare benefits.
The discrepancy between AI-driven welfare improvements and GDP measurements poses challenges for public finance, as government revenues depend on nominal GDP to service debts and fund expenditures. This gap highlights the difficulty in addressing political disputes over resource allocation, even as AI reshapes societal well-being.
In addition to benefits, AI introduces risks such as the proliferation of deepfakes, online scams, and mental health impacts on adolescents, all of which remain largely invisible in economic statistics. Inequality is another pressing issue, as AI tends to favor highly skilled individuals—including scientists, executives, and lawyers—while potentially displacing average-skilled workers. Some argue that falling costs of many goods and services might offset stagnant wages; however, this view may underestimate the growing share of income that individuals must spend on scarce resources, particularly housing in desirable urban locations. Since AI primarily augments intangible and location-independent attributes like intelligence, it is unlikely to increase the supply of sought-after properties, pushing real estate prices higher relative to incomes.
Consequently, concerns about inequality may intensify rather than diminish. Reduced prices for digital goods and manufactured items are unlikely to compensate for limited opportunities faced by many earners unable to afford housing in major cities such as London, Paris, Seoul, or Shanghai.
Ultimately, while AI has the potential to both improve and undermine aspects of human welfare, the overall outcome will largely depend on policy responses. Many of AI’s most significant effects, positive or negative, may remain hidden from conventional economic indicators like GDP, underscoring the complexity of navigating this transformative era.