“You have just been laid off by an algorithm”
Will such a statement become widely used in firms, or will it stay confined to grim dystopias?
In any case, it certainly reflects a growing trend in tech: the use of cost-cutting algorithms in human resources (HR), a new phase in the ongoing restructuring of this sector. The practice is part of an economic and geopolitical context that prompts most firms to adjust their financial outlook to ever-increasing demands in speed and performance.
Incidents such as Meta’s lay-off of sixty contractors that an algorithm picked at random, or the dismissal of 150 employees deemed ‘unengaged and unproductive’ at Russian software firm Xsolla, are alarming. Expeditious decisions on this scale are worrying, in themselves. However, what is especially controversial is the very use of algorithms to determine whether an employee is efficient enough to stay on the payroll. Why?
The specific challenges of using algorithms in HR
Using algorithms in HR creates challenges that are different to those in other fields, given the social impact HR practices can have. The most famous example of this is Amazon’s recruiting algorithm, which over time was rendered sexist by male predominance in the job performance data that it used.
But beyond tech giants, the scope in which HR algorithms can be applied remains limited. This is due to two major restrictions:
- The first is technical.
You need massive volumes of data to yield conclusive results. To provide relevant results, an algorithm needs to process data collected over time about individuals’ recruitment ( those recruited versus those that were not) and efficiency ( those who were most efficient and those who were not). The problematic here is that many organisations have limited usable data to allow for more sophisticated analysis. First, they do not keep data about candidates that they did not recruit. Second, they have too little data about their actual employees. While other categories of data can be used (internal or external, HR or otherwise) to make up for this lack, this raises new issues such as matters of governance and accreditation, for example.
- The second restriction is ethical.
Decisions in HR directly affect individuals particularly because they touch on broader social issues like equality and fairness in recruitment. Regulatory frameworks, such as the duty of care legislation in France, steer recruiters towards fair decision-making and hold them accountable for the choices they make.
A modernisation driver nonetheless
The rise in new techniques should still help modernise HR departments, given the limits they face. A first step is to look at algorithms in terms of their commercial and innovation potential, not just in terms of cost.
Algorithms give HR departments the chance to bring about human and social change – as long as this change is managed with a vision that is meaningful and can be put into practice. In the past, the purpose of tech-driven organisational changes was to boost productivity by automating or optimising tasks. In modern times, however, incorporating algorithms is part of a different strategic approach that seeks to improve employees’ performance and commitment. Algorithms now represent a way to take the individual into account and help change the role of HR players, which has long been based on the impersonal or segmented management of employees.
The main question: how do we want to develop HR?
To think about how to best use algorithms in HR means letting players in HR keep some control over the production and processing of their data, and allowing them to use their own discernment about possible consequences. The lifecycle of data (from data production to data processing) is a non-linear process that requires a considerable amount of collaborative, discretionary and situated work. Therefore, the acculturation and training of HR actors in data analysis is an essential element in making them aware of the use of algorithms in this field.
This post gives the views of its author, not the position of ESCP Business School.