Master in Decision and Innovation
First at alI I would like to congratulate all of you for your participation in the activity and hope you have enjoyed as much as we did. As you know, all of us are immersed in a reality progressively taken over by AI.
In this activity related to the Human Resource Role in companies we wanted you to reflect on how AI can help HR to predict your personality and skills. We know this possibility raises unsettling scenarios akin to “black mirrror” plots.
So our first concern was to assuage any alarm AI’s potential could bias your performance of the experiment we wanted you to take part. Hence we asked you to submit, -without saying what was it for- at least, 2000 words in which, you would explain who you are, your hobbies, what aspects define you, how do you approach life, etc…
As you already know after doing the experiment you have been able to compare what IBM Watson Personality Insights program thinks about your “real profile” taking into account the analysis of your writing description within the Big Five model.
Beforehand, we provided you with extensive information on how the algorithm works on the Human Resource Role activity’s page 3 on www.gaianext.com
As IBM Watson recommends in order to obtain more precise results we advised you to write the 2000 words text in Spanish or in your mother tongue. Sorry if we misled some people into writing in Catalan, Basque or Galician because currently the program only analyses in English and Spanish.
Our intention was to introduce you through a scientific approach to the basis of IBM Watson’s fundamentals backing the AI´s IBM software. To support the experiment we invited Luis Eduardo Imbernon to lecture on AI’s potential for recruiting based on his PhD research on AI and education.
Before Imbernon’s webinar, students were asked two questions:
- Do you think that AI could really predict people’s behaviour and preferences?
71,1% thought that it’s possible
- Do you think that HR departments could take advantage of those predictions to face a recruitment process?
83% thought that it’s possible
After the webinar and the students experiment themselves with the IBM Watson Personality Insights, we asked two similar questions:
- Do you think that IBM Watson has provided you correct results about your personality using your 2000 words text?
76% thought IBM’s program was correct in its prediction.
- Do you think that this service performs better the work of guessing a candidate’s personality than a professional recruiter?
19% thought that IBM Watson perform better than a recruiter.
It’s clear that students’ opinions about AI’s power prediction on personality didn’t suffer significant variation after trying the software. On the other hand, regarding the second topic (AI’s ability to improve human recruitment) has suffered a significant decrease (64%) in the support students gave it in first instance.
It is true that the second question wording had probably pushed many students to reconsider their former sympathetic opinions towards AI potential. However, in light of this decrease of support for AI’ use in recruitment we wonder if students’ opinion were also strongly influenced by theirs visualizing a not-so-distant scenario where machines may effectively assess humans without taking into account human interaction, feelings and the ever present aspiration that chance and fortuitous meetings matter.
Welcome to the near future! Enjoy it!