The use of recruiting algorithms to search for and identify the best candidate from billions of data is not a new concept. However, few companies in need of recruitment and retention have been encouraged to introduce them into their processes, at least in their early stages. Maybe it is because the fairness they promised may not be so fair.
“The automation bias”
Between 2014 and 2017 Amazon developed a system of algorithms to analyse profiles and detect the best ones. It promised, and a lot. But after the tests, it became clear that the algorithm left many people away from the desk of the HR manager, running the risk of discrimination. Amazon was not the only one, but it did open the debate about these systems.
“It is not true that they are neutral (algorithms), because they learn and are based on reality, which can be biased. And the fact that someone has programmed this algorithm means that it is possible that they have overruled their prejudices,” says Adrián Todolí, professor of Labour Law at the University of Valencia in this article.
It is a phenomenon known as “automation bias”. With the latest advances, machines learn better and faster, but humans program them, teach them. And human beings are biased, whether conscious or unconscious of it.
“The great barrier of automation”
Artificial intelligence and automation have enormous potential, especially in the early stages of identifying the right data. But the key is to have a large capacity for relevant information. And not only that, the quality of the data. The greater the volume of data, the smaller the possible discriminations will be. But what information is relevant? Is it relevant that they know about my daily behaviour? From my travels through photos posted on Facebook or Instagram? My opinion about a restaurant five years ago? We could get into quicksand. Asking ourselves the right questions is key.
In addition to the great challenge of having relevant information and managing it to obtain ideal data, we continue to encounter the problem of bias. And bias is not something that is originally in the algorithm, but that is transmitted by humans. And the human factor continues to be a determining factor in decision-making. There are variables that cannot be substituted by automation when assessing whether that candidate is the ideal person: cultural, organisational fit, or chemistry with the team. Organisational psychology still has a lot of work ahead.
It is evident that the use of algorithms plays a fundamental role in each of the phases of the recruitment process and in each of the decisions taken in it. Especially in the first phase, which is identification. It has been shown to be effective and really help in decision making by providing objectivity and consistency. But algorithms also carry their own risks, even when they eliminate subjectivity from recruitment processes. In the end, people are involved in the programming of those automation systems and in the final decision making.
“If a company is analysing profiles based on algorithms, the hiring manager needs to be sure that the algorithm is valid,” says Heather Morgan, global chair of the workforce data and technology practice at law firm Paul Hastings in this article. “You have to ask the right questions. Because it is difficult to validate something that is continually changing and evolving.”
“Data-based technology applied to the best database: the Web”
Nowadays, the most suitable professionals can be anywhere in the world, their identification requires the application of smart technologies with global reach and in constant updating.
Our Sourcing Innovation team uses the latest technology and techniques to discover hard-to-find talent in the more than 800 million professional items indexed on the Internet using elements such as employer branding, boolean search, social media hunting, data analytics and our own collaborative models and tools such as Social Referral or Talent Hackers, and AI solutions such as our proprietary KM Crawler among others
Catenon benefits from the expertise of its innovation hubs in order to incorporate pioneering and innovative data technologies into our processes to efficiently identify candidates worldwide.