Various studies have considered the impact of automation on jobs. The latest report from Centre for Cities is broadly consistent with these, stating that c. 20% of roles are highly likely to shrink by 2030 (that’s around 300,000 per annum). The cities most vulnerable to this displacement are typically mid-sized with high exposures to sales, logistics and administrative roles. More positively, the report suggests that all cities will, in fact, benefit from a net addition of jobs over the same period, once newly created roles are factored in. In support of this, the report highlights the destruction of historic roles such as domestic servants, which were since replaced by new ones. To me, this misses the bigger point. Reskilling is a part of every industrial change, and migration to more productive occupations is the desired outcome. However, what happens when there is no job below a certain skill level that cannot be performed more efficiently by a machine? Machines are getting smarter every year, whereas humans are not, and so this outcome feels like an inevitability sooner or (hopefully) later.