Is AI a staff-cutting tool? IBM's own automation chief says using it that way is 'wrong'

Neel Sundaresan, IBM's general manager for automation and AI, told the Chosun Ilbo in Seoul that AI is 'not a tool for cutting staff but a tool for creating more value with the same workforce', and named the reasoning he rejects: approaches like 'productivity has increased by 30%, so we can reduce staff by 25%' are, in his words, wrong. He put numbers behind the claim. About 163,000 of IBM's 270,000 employees now use its internal assistant, which writes roughly 40% of its own code, and the company measures an average productivity gain of 45% with no decline in code rejection rates, bug rates or weekly completed tasks. Junior engineers use it as a senior, he said; senior engineers use it as a junior. The interest here is the tension inside one company. IBM supplies this site's clearest case of literal replacement, a named system taking 200 named HR roles, and now its automation chief arguing that treating a productivity gain as a headcount cut is a misuse of the tool.