The information hit the headlines last September, NetDragon Websoft, a Chinese technology company specializing in online games and the metaverse, appointed a humanoid robot equipped with artificial intelligence as CEO. Called Tang Yu, he will serve as ” real-time data center and analytical tool ” At the direction.
AI cannot totally replace managers
In an exciting column published on The ConversationRobert Donoghue, researcher at the University of Bath, and Tiago Vieira of the European University Institute, returned to this growing trend within companies to use algorithms to carry out management tasks that were previously devolved to managers.
This ranges from the examination of applications, to the evaluation of employee performance, through the decision whether or not to dismiss an employee. Managers are often seduced by these tools, which reduce costs by automating certain tasks. The authors cite the example of Uber, which supervises its 3.5 million drivers using technological devices.
But these systems are clearly not without risk. The researchers mention the problems with artificial intelligence biases. The example cited is the resume ranking system implemented by Amazon. The tool had very clear gender biases and it was soon abandoned.
Machine learning algorithms also raise questions. While their designers are sometimes unable to understand how they work, the lack of transparency is glaring. How then imagine entrusting the future of employees, even their dismissal if one is unable to say why a decision was taken?
The authors also believe that an artificial intelligence cannot logically demonstrate humanity. In a study they conducted of Amazon Flex delivery workers, many employees complained about his often ruthless decisions.
Thus, the machine is not able to understand that a worker is less efficient because he has just started his job or because he is sick or has a disability. Likewise, they systematically refuse or sanction any deviation that could be justified: an unexpected departure for a family emergency, for example.
On the strength of these lessons, the researchers provide some ideas: “ US politicians are discussing an extension of digital rights for workers. Other solutions include regularly evaluating the impact of algorithms on employees and giving them a say in how these technologies are used. »