Artificial intelligence promises to reshape economies by opening new channels for innovation, manufacturing, education and problem solving, yet it also risks widening existing gaps in wealth and worsening environmental pressures if left unchecked. Integrating AI-driven progress with social goals that prioritize shared prosperity, justice and ecological balance, according to Alessandro Crimi, presents a formidable challenge for policymakers.
Historical experience shows that merely retraining workers does not automatically raise living standards. In medieval societies, guilds, churches and monasteries provided informal apprenticeships, and new machines such as windmills created demand for millwrights. Nevertheless, the era is remembered as bleak because productivity gains failed to translate into broad material improvement, suggesting that profit-sharing, taxation and safety-net measures are also required.
Retraining initiatives remain essential but insufficient, as they shift the burden of adjustment onto individuals and often lag behind the speed of technological change. Complementary structural policies,such as reduced working hours, a four-day workweek, or universal basic income,aim to preserve economic security regardless of employment status. Yet financing these programs raises the question of where the necessary resources will originate.
One proposal to generate those resources is an automation impact levy, often called a robot tax. The idea targets a market failure in which firms capture wage-bill savings from automation while society bears the costs of unemployment and community decline. Early experimental work indicates that such a levy can lower the likelihood that firms replace workers with machines.
Designing the levy, however, faces practical and philosophical obstacles. Determining whether the taxable unit is a physical robot or an AI algorithm proves difficult because automation often involves integrated software rather than discrete hardware. Prominent tax scholars have suggested that reforming broader capital taxes may be more effective than a narrowly targeted robot levy, arguing that its appeal may stem more from behavioural bias than sound fiscal logic.
Real-world attempts illustrate both the promise and the resistance to such measures. In 2017 South Korea reduced the tax credits that large and mid-size firms could claim for automation equipment, effectively withdrawing a subsidy rather than imposing a new tax. The same year the European Parliament rejected a robot-tax proposal, with many legislators fearing it would stifle growth and innovation, underscoring the political difficulty of taxing machines.