Cracking the Hiring Algorithm

As hiring interviews go the AI way, cracking them is proving to be a tough puzzle even for data scientists

It’s a unique irony of fate that data scientists, are getting frustrated when faced with artificial intelligence (AI) chatbots created by data scientists like themselves. As Human Resources gets upended by technology with organizations using AI for a range of activities starting from trawling social media to identify and profile candidates, to shortlisting CVs, scheduling interviews and even using it to conduct job interviews, job seekers are grappling with the need to adapt to chat-bot interviews; and often, it has become a stumbling block for many.

Job seekers, which includes even data scientists, are missing the human touch, as the world moves to a remote-everything contactless economy, forced by the pandemic. Interviewees are also unable to keep pace at which the AI chat bots are asking rapid-fire questions and complain of not been given enough time to think. Clearly thinking on their feet is something that candidates must learn if they wish to crack the code of the algorithmic interviewer.

There’s no option for jobseekers but to get used to having their CVs selected and themselves interviewed by an algorithm; and it is tough. According to Jobscan, a US web service that helps job seekers land more interviews by using AI to analyse one’s resume or LinkedIn profile against any job description, more than 95% of Fortune 500 companies use ATS (application-tracking-system). Yet 63% of candidates are dissatisfied with the way employers communicate.

Those being interviewed by an algorithm are getting flummoxed by the speed and the ability of the chat bots to throw unexpected challenges like writing a piece of code within a given time. “There is no room to negotiate for an extra few seconds with an algorithm,” bemoans a candidate. “I do not know what kind of facial expression would be liked by the chat bot,” wails another job-seeker who has failed to crack a chat-bot interview, while someone else thinks that it might be easy to game an algorithm by faking expressions to appear smart and intelligent.

According to the 2019 Artificial Intelligence in the Recruiting Industry benchmark report of CEIPAL, SaaS provider of staffing companies – two-thirds of all staffing firms will adopt AI-driven ATS by the end of 2020, while 79 % of enterprise staffing firms (firms with more than 100 recruiters) will have done the same. Gartner’s latest survey found that 23% of organizations who were already piloting or using AI, were doing so in the HR and recruiting domain.

Hiring organizations on the other hand are rather happy with using AI to make the HR process faster and efficient by taking away tedious labour-intensive tasks like sorting CVs and getting the basic information about a candidate, specially pre-screening potential employees. Companies also feel that algorithms are free of human biases (a point that can be debated) and is fair to jobseekers.

After the first step of screening is complete, interview chatbots can move on to the qualification process by identifying the strengths and weaknesses of the candidates, analysing the skills, qualifications, and past experiences included in résumés and cover letters. Recruitment bots therefore save significant time and effort for human recruiters before setting up face-to-face interactions. The process of attracting, screening, and interviewing candidates for a job seems deeply human, but certain chatbots can detect and register conversation emotions. Developed platforms can understand a user’s mood throughout the chatbot interview process and can score sentiment based on connotation, word placement, and grammatical modifiers.

Those who have cracked the hiring algorithm feel that as the program uses visual analytics for reading facial expressions it is important to always look at the camera so that eye contact is never lost. The way to handle the situation is to pretend that one is speaking with a live person, and act accordingly, and make sure to enunciate your words and try not to stutter or have long pauses.

HireVue, which is one of the most popular AI-based hiring platforms used by several Fortune-500 companies have a candidate help section in the website for practice questions, recording interviews by the job-seeker themselves to evaluate their own performance and make adjustments before appearing for the actual interview. HireVue’s standard questions test out the applicant’s ability to handle crisis situations, project management skills, adaptability, teaming capabilities, customer relationship management, negotiation skills among others.

Candidates are being evaluated on their E-Charisma as well which is the personality, they project in collaboration platform meetings as in like Zoom conference calls. Here again eye-contact, smiling pleasantly whenever required are some of the non-verbal cues that are assessed. Nodding in agreement, or making noises like “uh-uh”, are irritants which fetch lower scores in the E-Charisma chart. Tone of voice, easy flow of conversation and the ability to draw others into the discussion also adds to the score.

A remote-everything contact-less world requires new skills to survive and thrive, especially soft skills. Along with articulation and a high level of communication capabilities, we now need to be able to learn how to appear relaxed, comfortable, and easy in front of the camera. One must remember that faking expressions don’t work. Being in front of a camera requires hard practice and a lot of homework. More than in the physical world, one is always being measured when one’s on the screen, as there’s nothing else to distract the viewers’ attention as was the case in a face-to-face meeting in the pre-Covid world. If you have to act, then act natural!