← All research

AI can improve productivity. Managers may not know where.

Only 7.8% of managers in the study correctly anticipate differing effects across tasks. Heterogeneous performance can then become a homogeneous expectation.

An average becomes a norm

The study distinguishes tasks on which AI markedly improves performance from others on which it degrades it. Yet only 7.8% of managers correctly anticipate this distinction, while 92.2% expect improvement across all tasks.

Organisational risk emerges when a heterogeneous technological effect becomes a homogeneous expectation: raised targets, shortened deadlines or mandatory use regardless of task and person.

Change the unit of analysis

Benefit does not depend only on task, model and worker. It also depends on organisation and deployment choices: training, permission to doubt, verification, workload and the capacity to report failure.

Better deployment means not only adapting workers to the tool, but adapting the organisation and use of the tool to real work.

03

Why it matters

If benefits are wrongly generalised, the organisation may try to correct the worker when the problem lies in the task, model or deployment.

04

Source

LinkedIn — analyse d’une étude organisationnelle

Open the source Originally published on LinkedIn