The 'ironies of automation' applied to AI at work
Can AI erode the skills it is meant to support? An old cognitive-engineering paper offers a useful frame.
The original argument
Lisanne Bainbridge (1983, Automatica) shows automation leaves operators the rarest, most complex situations while they practise the task less. The paper is conceptual, from industrial control.
Transfer to AI
Raja Parasuraman and Dietrich Manzey (2010, Human Factors) document reduced vigilance with automated aids. For generative AI, evidence on skill loss is still limited.
The nuance
The point is not to refuse the tool but to deliberately preserve practice of skills humans need when it fails. (our executive and employee training programmes)
The limit of common practice
Organisations measure time saved, rarely residual ability to do the task without the tool.
A practical check in Bern
Once a month, have a usually assisted task done without AI and compare quality and duration. Gradual decline signals dependency risk. To go further: explore the AI on the Job training in Bern, or browse our executive and employee training programmes in Switzerland.
In pictures: AI on the Job in Bern
- ai on the job
- Bern
- research
- training bern