Openjobmetis: horizons and limits of innovation

Let’s start from the context: how does Openjobmetis come into contact with artificial intelligence?
Openjobmetis is a job agency that deals with human resources management in a comprehensive way: search, selection, training, hiring and outplacement. I believe that the human resources sector is currently deeply involved in reflecting on the meaning of introducing technological innovation tools, especially in the pre-screening phase of candidates, where it is important to have systemic support in the processes of initial analysis and routing of the many applications we receive every day. Our pre-selection process thus finds in artificial intelligence a strong and integrated ally for the initial profiling of CVs in the database.

How do AI agents facilitate the work of recruiters?
Let’s consider that we receive over 30,000 CVs per week: a significant volume of applications that makes it clear how, if we had to analyze them one by one, the process would require an excessively time-consuming effort, whereas at this stage what is needed is to select CVs effectively based initially on criteria of compatibility with the proposed job description.

And what about the selection phase?
In the actual selection phase of candidates, for now we continue to prefer in-person interviews.

Do you think AI can overcome the problem of a priori discrimination?
I would start by reflecting on the fact that bias toward our interlocutor is sometimes unconscious and therefore also present in professionals involved in selection. We all have cognitive biases, and in this sense an artificial intelligence tool such as an avatar could, at least in theory, adopt a more objective approach. However, we must take into account that this kind of practice and adoption is still experimental for us: we do not rule anything out a priori, but naturally we must verify how much the system can be trained to provide real support. For example, we are evaluating features capable of detecting how comfortable our interlocutor may be during the dialogue, progressively adjusting the interaction; not to mention that some systems also provide accessibility features for people with physical difficulties, such as speech recognition or tools that allow expression through sign language for deaf people. It is a world that until ten years ago seemed impossible and that today instead opens many doors, because accessibility also means neutrality. Today’s labor market, characterized as it is by a constant need for personnel and the difficulty of finding it, forces even the most reluctant companies to question themselves in the selection process. Even very small organizations are working to make their external communication more engaging, more attractive and able to show potential candidates their strengths.

Could there be risks in AI-integrated selection processes? Which ones?
Regarding whether virtual agents can be fully effective, we must proceed on the one hand with an open mind, and on the other always with conscientiousness. The experience of candidates is a distinctive element for us, because they are an extremely valuable asset. It is essential for us that candidates trust us; therefore, we must be credible in their eyes from the very beginning, and we can only achieve this through high-quality human interaction. Beyond this, there is another major risk I see: a countervailing trend toward possible social exclusion. Excessive digitalization and automation can exclude entire segments of the population that are not sufficiently digitally literate. This is an important issue, especially in a labor market that must deal with a progressively aging population.