ABSTRACT

Progress in AI is bound to take place in fits, starts, and leaps as it is based on research characterized by progress, which is definitely not smooth. Thus, there might be a clustering of adoption of AI in different sectors at points of time, which in turn would give rise to bulges in unemployment. Given that such bulges hurt numerous workers and are economically unjust as well as give rise to recessionary tendencies across broad swathes in the economy, there is a case for generating a queue for adoption of AI through high taxes – on employment of robots in some of the economic sectors at a given point of time to prevent excessive clustering – which would eventually get scaled down in the future to make robotization economically viable. The tax revenues generated can be used for financing minimum incomes for the unemployed, retraining them for new jobs and helping them acquire financial and physical capital with associated returns offsetting the risk in regard to returns emanating from human capital.

It is very important that these taxes are administered on a global rather than national basis; taxes at the sectoral level, which differ across nations, are a trigger for capital flight to low-tax destinations and ultimately result in competition between nations, which reduces tax rates to zero.