Will generative AI bring the productivity increases we need?

August 1, 2023

Many worry that AI will have a significant impact on jobs across the economy. From call center workers to business analysts, nobody can be certain that his or her job won’t be displaced by intelligent systems. While this may be worrisome from an individual perspective, it could also provoke a sigh of relief; replacing human workers with machines could mean that we finally see some productivity increases after decades of computerization and digitalization.

However, despite the apparent advancements in information technology, there has been a lack of corresponding increases in productivity. Various explanations for this productivity paradox exist; it may take more time for productivity effects to become visible (as businesses need time to fully implement digital solutions), it could be a measuring problem (traditional statistics miss the value generated by digital technology), or perhaps digitalization has primarily created new practices that don’t truly add value (e.g., elaborate PowerPoint slides, superfluous emails). 

For generative AI, the question is whether it will reduce our workload and lead to time savings (and hence productivity increases). Alternatively, whether the lower costs to produce text, images, data analyses, and other outputs that AI’s can deliver will inspire new nonsensical practices, such as even more intricate slides or personalized services that barely add value. Moreover, if AI does indeed save us time, will it do so in the sectors where it is needed most?

Chances are that AI will primarily replace workers who are of minimal economic and societal value. These so-called 'bullshit jobs' often entail ‘information work’, which is precisely what AI excels at. By contrast, substituting people in sectors that hold high societal value and currently struggle with labor supplies will be far more difficult. Jobs in healthcare, education, and construction typically involve combinations of cognitive and physical tasks, which are much more challenging to automate.

These likely outcomes suggest that we could witness a growing number of unemployed ‘information workers’, while other sectors will continue to struggle to find skilled workers. From a societal perspective, this raises the question of whether and how we can transition people out of their current 'bullshit jobs' and into sectors that are of crucial importance to society and the economy.

Burning questions:
  1. As generative AI continues to evolve and potentially displace workers, how should society and governments adapt their educational systems and training programs to ensure that individuals are prepared for sectors where human labor is still vital and in high demand?
  2. How can we foster innovation in AI that delivers productivity increases even in hard-to-automate sectors? For instance by reducing the administrative burden on healthcare workers or teachers?

About the author(s)

Fascinated by the interplay between technology and society, Sjoerd has studied the role of different actors in the innovation and implementation of new technologies throughout his career. At the thinktank, he is mainly involved in research and consultancy projects for clients, and strategic and thematic research for sister company Dasym. Among other themes, Sjoerd frequently writes and speaks about the power and danger of digital technology, as well as sustainability in both technological and institutional innovation.

You may also like