copyright by www.theglobeandmail.com
As we know, they finally showed up. But they took longer than expected because they depended on investments in complements.
As in the computer age, the widespread productivity gains associated with artificial intelligence will depend on investments in complements, which are all the things other than algorithms and models needed to make commercial-grade AI work. These include data, redesigned workflows, training, regulation, human judgment, infrastructure and more.
Paul David, a Stanford economist, responded to Prof. Solow’s productivity paradox remark by examining the history of electrification after 1900. Mr. David’s key insight was that the productivity gains took a surprisingly long time after the initial invention – more than two decades – due to the many other complementary investments required to fully realize electricity’s benefits. For example, distributed electricity allowed for powered machinery to be spread out rather than stacked to enhance material handling and workflow, but to do this required factory layouts to be entirely redesigned as single-storey rather than multistorey. And electric-powered machinery could be made more modular so that downtime for one machine did not require downtime for all, but that also required redesign.
Mr. David wrote: “At the turn of the century, farsighted engineers already had envisaged profound transformations that electrification would bring to factories, stores, and homes. But the materialization of such visions hardly was imminent. In 1899 in the United States, electric lighting was being used in a mere 3 percent of all residences (and in only 8 percent of urban dwelling units); the horsepower capacity of all (primary and secondary) electric motors installed in manufacturing establishments in the country represented less than 5 percent of factory mechanical drive. It would take another two decades, roughly speaking, for these aggregate measures of the extent of electrification to attain the 50 percent diffusion level.”
As we shift from technical achievements in AI (“Look everyone – the AI can read the address on an envelope! The AI can classify a medical image! The AI can play a video game!”) to large-scale commercial deployment, the design and implementation of complements will be paramount.[…]