Development of Industry 4.0

“Smart factories” are being designed where cyber-physical systems monitor the
physical processes of the factory and make decentralised decisions.
To demonstrate, a leading robot manufacturer, Fanuc, recently
unveiled an assembly-line robot which can learn to pick randomly-
orientated objects out of a bin. A video camera picks up footage of each
attempt and feeds the subsequent success or failure back into the system
for use in the next process. And this learning can easily be shared. In
other words, five robots working together for one hour can accomplish
what one robot can do in five hours.

The same scalability does not apply
to human workers who cannot share learning quite so efficiently.
Another example is a loading robot. After four hours of training,
the success rate of part transfers rises from 50 per cent to 70 per cent.
After ten hours, it rises to 90 per cent, close to the level of a skilled human
worker.

Finally, artificial intelligence-based breakdown diagnostics can
detect problems in a system up to 40 days in advance of a problem. In
contrast, humans were only able to detect a problem when it became
noticeable that it was about to break. Some researchers estimate these
types of intelligent systems may replace six per cent of US jobs by 2021.