The question of whether these systems replace people or make them more productive, which usually has different answers for different tasks within the same job. Most roles are bundles of tasks, and automation tends to take some and leave others, which changes what the job is rather than removing it outright. The honest position is that effects so far are uneven and the measured evidence is thin relative to the volume of prediction.
Technical view
Task-level exposure analyses consistently find that a substantial share of tasks in many occupations is affected, while noting that exposure is not the same as automation, since complementarity, regulation, liability and organisational inertia all mediate the outcome. Early field studies show the largest productivity gains going to less experienced workers, which compresses performance distributions and has implications for career progression. Historical precedent cuts both ways: technologies have generally raised aggregate employment while eliminating specific occupations and imposing concentrated costs on the people in them.
Related
Part of AI Terminology Index, section: Society and Impact.