The world’s first robot designed to carry out unbiased job interviews is being tested by Swedish recruiters. But can it really do a better job than humans? Her name is Tengai. Measuring 41cm (16in) tall and weighing 3.5kg (7.7lbs) she’s at eye level as she sits on top of a table directly across from the candidate she’s about to interview.
Her glowing yellow face tilts slightly to the side. Then she blinks and smiles lightly as she poses her first question: “Have you ever been interviewed by a robot before?” Tengai is the brainchild of Furhat Robotics, an artificial intelligence (AI) and social robotics company born out of a research project at Stockholm’s KTH Royal Institute of Technology.
The firm has spent the past four years building a human-like computer interface that mimics the way we speak, as well as our subtle facial expressions. The idea, according to chief scientist Gabriel Skantze, is that “it feels much less scary or strange compared to a more traditional robot”.
Since October 2018, the start-up’s been collaborating with one of Sweden’s largest recruitment firms, TNG. The goal is to offer candidates job interviews that are free from any of the unconscious biases that managers and recruiters can often bring to the hiring process, while still making the experience “seem human”.
“It typically takes about seven seconds for someone to make a first impression and about five to 15 minutes for a recruiter to make a decision. We want to challenge that,” explains Elin Öberg Mårtenzon, chief innovation officer at TNG’s office in central Stockholm.
Unconscious biases include making assumptions about someone’s competence based on gender, ethnicity, voice, education, appearance, or as a result of informal conversations before or after an interview. “For example, if I ask you a question at the beginning of the process like: ‘Do you play golf?’ and you say: ‘Yes I do, I love playing golf’, and I do too, then in some way I will put that in a positive box,” says Ms Mårtenzon.
Tengai, by contrast, doesn’t engage in pre-interview chit-chat and poses all questions in an identical way, in the same tone, and typically, in the same order. This is thought to create a fairer and more objective interview. Recruiters or managers are then given text transcripts of each interview to help them decide which candidates should move to the next stage of the process, based on the answers alone.
“I think it is a really good help while screening – if you have a big recruitment process with a lot of candidates – having someone, a robot, with no emotions, no feelings,” says Petra Elisson, a 50-year-old who works in healthcare recruitment.
She has been taking part in the trials, both as a recruiter analysing candidates transcripts and as an interviewee. “I was quite sceptical at first before meeting Tengai, but after the meeting I was absolutely struck,” she reflects. “At first I really, really felt it was a robot, but when going more deeply into the interview I totally forgot that she’s not human.”
But robots are programmed by humans, analyse datasets compiled by humans, and learn from human behaviour, so couldn’t they inherit some of our biases? Furhat Robotics says it has sought to stop this by carrying out multiple test interviews using a diverse pool of volunteers.
“It’s learning from several different recruiters so it doesn’t pick up the specific behaviour of one recruiter,” explains the start-up’s chief scientist, Gabriel Skantze.