Amazon uses automated, AI-driven processes to monitor and evaluate the work of its employees and, if necessary, to initiate consequences. A report from the company shows that even warnings and notices of termination are sometimes sent automatically.
The basis of the report from The Verge is a letter from one of Amazon's lawyers. It reports that especially in Amazon's US fulfillment centers, the productivity of the workers is tracked and evaluated by a system. If a facility's employees do not meet their quotas, the system writes not only a warning to the superior or a message to the employee, but in some cases it can even directly issue the warning and termination completely on its own.
One of the parameters tracked is called "time off task", which indicates how much time elapses between two recorded scans by an employee. If this time, abbreviated TOT, is too long too often, there are problems. This means that employees avoid taking breaks and even avoid going to the bathroom just to stay in the system's good graces.
It is true that the superior of the employees concerned can veto and thus prevent firings initiated by the AI. But let's be honest: How often would a superior go out of their way to contradict the AI, especially if it concerns a position with low qualification requirements, which can be filled again relatively easily?
The figures in the document speak of 300 terminations due to inefficiency in one year at the Baltimore fulfillment center described in the report - around 10 percent of the workforce. Whether it looks the same elsewhere, is not yet known.
Don't push it!
Practices like the ones that have just been uncovered about Amazon can mean a dangerous path - for society as a whole. In many countries, including the USA, the gap between salaries is already far too wide anyway. As a result, the bond between employee and company is becoming more and more relaxed - the social structure of the "workplace" is crumbling more and more. Above all, however, people who earn little money depend on a certain level of security.
"What we hear again and again from the workers is that they are treated like robots, which is also understandable in view of the monitoring and control by automatic systems," says Amazon critic Stacy Mitchell of the Institute for Local Self-Reliance.
The fact that modern technology is used for optimization in the corporate environment seems good and right, and of course Amazon's AI probably isn't throwing anyone out just because a package was packed too slowly one time. What really bugs me is the fact that the human superior has to intervene in order to prevent dismissal - it should be the other way around (the AI should share its objective view of an employee's productivity with the superior when he or she is doing evaluations).
Do you think it's okay that employees can be fired by an algorithm? Let us know in the comments below.