Reorganisation not replacement

Director of the Stanford University Digital Economy Lab Erik Brynjolfsson argues that AI's impact on the workforce will be about restructuring work rather than mass unemployment.

At this point none of these technologies can do everything, but many will impact some tasks considerably while other tasks won't be affected at all. Leaders must recognise that into the future implementing new technology will be much more about restructuring and reorganising work rather than mass replacement or unemployment.

Brynjolfsson’s research that examined 18,000 tasks suggests that 50-60% of the workforce will be impacted by AI, but this doesn't mean widespread job losses. Using radiologists as an example, whilst AI can read medical images better than humans, radiologists perform roughly 30 distinct tasks - only one of which is reading images.

Quote: Erik Brynjolfsson Atlantic Interview 

Brynjolfsson's recent call centre research found that AI can actually reduce inequality. When large language models were introduced to assist operators, less-skilled workers saw a 35% increase in productivity whilst the most-skilled workers benefited almost zero. The AI captured tacit knowledge from experienced workers and transferred it efficiently to newcomers.

He believes that a key insight from technological advancement so far is that whilst technology makes specific tasks more efficient, it rarely replaces entire occupations. When costs fall due to increased efficiency, demand often increases even more - leading to greater overall employment in transformed roles.

Source: Thompson, N. 2025. The Jobs Equation– Erik Brynjolfsson. The Atlantic.