How Artificial Intelligence in Recruiting Wins War for Talents
AI, Machine Learning and chatbots help remote HR management in times of COVID-19
The coronavirus has forced companies worldwide to send their workforce home. Job interviews are being cancelled, some have stopped recruiting altogether. Did you know there are companies who have already placed their hiring process exclusively online? AI can help.
Digitalisation and globalisation have led to a fundamental change in the world of work. HR professionals are adapting to completely new professional titles, such as Big Data Scientists, Conversational Designers and AI Ethics specialists.
Artificial intelligence (AI) is already an integral part of the job application process in large companies and recruitment agencies worldwide. Meanwhile, Germany is once again lagging behind. In the following I will give you an insight into how AI applications can already support HR professionals today:
- AI and Chatbots: A Billion Dollar Market
- Sourcing, Screening and Candidate Experience with AI and Chatbots
- AI-driven online job interviews: Smile, please!
But first of all, let’s see what artificial intelligence is and what it is not.
Artificial Intelligence: Not a Myth and No Mystery
Some people still think of the Terminator or Sophia, the telegenic robot girl when you mention artificial intelligence. Or they picture an omnipotent computer program that has surpassed human intelligence and gained self-consciousness (this is actually still pure science fiction and is called General AI or strong AI). AI is also often equated with machine learning, but ML is only one part of AI.
Wikipedia defines AI as “In computer science, artificial intelligence (AI), sometimes called machine intelligence, is intelligence demonstrated by machines, in contrast to the natural intelligence displayed by humans.”
Put simply, AI applications are learning software that simulates intelligent behaviour.
How sophisticated, that depends on the application. Present-day AI applications are called weak or narrow AI, which is task-related AI. This type of artificial intelligence is behind automatic image recognition (already helping doctors to detect breast cancer at an early stage), making the Tesla and Google self-driving cars possible and helping police officers search for criminals, using facial recognition. Siri, Alexa and every half-decent chatbot use natural language processing (NLP) to understand the many nuances of the human language — that too is AI.
AI in Organizations: A Billion Dollar Market. And One in Four Germans Want to Speak to a Chatbot
The global chatbot market will grow to 9.4 billion US dollars by 2024, the European AI market will increase fivefold to 10 billion euros within five years, and already one in four notoriously cautious Germans wants to talk to a chatbot.
Chatbots, voice assistants and other dialogue-based AI systems are changing the way we deal with information in all areas (“Hey Siri, add matches to my shopping list, please!”) and so-called “robot recruiting” is already playing an important role in international HR departments.
Germany’s HR has to step up and look beyond its national borders if it is not to lose out in the battle for talent: to date, only 3.9% of German companies use data-driven tools in their recruitment process, although 44% see them as a useful addition to traditional procedures (IFO Institute, 6/2019).
AI solutions for the following HR areas are particularly interesting:
- Optimize job advertisements
- Sourcing (pro-active talent acquisition)
- First contact through chatbots
- Screening of applicants by digital assistants
- Video interviewing using AI
- Assessment
Professional AI recruitment platforms now cover the entire life cycle of personnel management. Here I will mainly focus on
- the advantages of initial contact and screening using Conversational AI (chatbots)
- Promises and dangers of AI-driven job interviews via video call in times of COVID-19
First Contact: A Chatbot. Day & Night.
Nowadays, the initial contact is made via a job portal or your company’s career site. Chatbots and voice assistants never sleep and can handle thousands of requests simultaneously. Baby boomers, even more than millennials, expect companies to be available online 24 hours a day and have high expectations for automated customer service. Questions about the company or the application process can be answered effortlessly by a chatbot on your website who never gets tired, impatient or sick. Or bored beyond belief by having to answer the same questions over and over again a thousandfold.
Trending AI Articles:
1. How to train a neural network to code by itself ?
2. Paper repro: “Learning to Learn by Gradient Descent by Gradient Descent”
3. Reinforcement Learning for Autonomous Vehicle Route Optimisation
Artificial intelligence, NLP to be precise, is behind every good chatbot these days who understands and answers questions in the form of natural language. This works with written text as well as with speech recognition — think Alexa, Siri & Co.
Chatbots in Recruiting at Work Everywhere
AI-based recruiting systems already increase the productivity of HR departments by 144% internationally (1). In Germany, Austria and Switzerland you can also find them in action: Chatbots and AI-supported services are already active in the recruiting process, for example, Allie from Allianz in Facebook Messenger, KATY, the career bot from T-Systems (2) or Jobpal at hub:raum, even if they come across rather primitively with buttons. Somewhat more sophisticated bots that can understand and process more complex issues can be developed with Mya.com (also available in German and many other languages). Mya specializes in AI applications for the entire HR life cycle.
The hiring process in many industries is already completely digitalized and candidates are used to answering questions online in job application portals and uploading all documents. However, if a company wants to position itself as innovative and reach out to the best talents, it must no longer scare off applicants with a five-page form. 87% of all applicants perceive this kind of technologisation of the application process as impersonal (3).
The better option: a friendly, tireless digital career assistant on the website and in Facebook Messenger is available for candidates where the questions arise — and that goes both ways.
Screening: AI-controlled Chatbots Excel in Doing the Most Boring Work Faster Than Anyone
For HR managers and recruiters, AI-driven bots are a dream come true: the virtual recruiting assistant, who “lives” on the website or Facebook Messenger, takes care of the most repetitive and boring parts of the job, such as screening applicants and scheduling interviews. This is traditionally attempted through endless questionnaires, but 74% of all applicants drop out in the middle of the application process (1) — probably just the talent you want!
A friendly, round-the-clock career bot, on the other hand, is attractive for job applicants and patiently answers the frequently asked questions day & night. This has already resulted in completion rates of 93% (2).
The bot can also arrange an appointment for a first interview if the candidate seems suitable. By connecting the bot to the company’s own calendar or ATS, the process can be completely automated and saves valuable time.
These AI-supported bots can not only read LinkedIn profiles and CVs but can also ask specific questions about gaps and experiences in a direct dialogue with the applicant. This is a classic but very repetitive recruitment activity that is now automatically completed in advance. It saves you almost 80% of the time from the initial contact to the interview. This is no longer science fiction: Such a dialogue can be text-based or with natural spoken language — just as you know it from talking to Siri, Alexa or your Google Assistant.
Platforms for the design and implementation of AI-supported chatbots that can communicate with people in natural language are now in abundance and there is something for every production budget. The best platforms also offer connectivity to your internal HR system, ATS or intranet, as well as seamless integration into existing work processes.
Job Interviews with a Robot? Smile, please!
Internet giants Apple, Amazon and Google, but also L’Oreal, Vodafone or Goldman Sachs (and more than 700 other large companies) go one step further: They even let the first interview be conducted completely alone by AI-supported software via video call with the candidates. Not always to the amusement of the candidates — to be interviewed by a faceless software and to smile as competently and enthusiastically as possible into your own mobile phone is not only unusual but also feels discriminatory.
HireVue is the market leader in this area and advertises with slogans such as “Beating Bias with AI”, but the technology is very demanding on applicants and is controversial, especially because of the emotional analysis using face recognition (more on this in a moment). ModernHire also offers such fully automated live interviews as part of its recruitment platform, which covers all areas of HR management.
During a fully automated structured interview, the software uses the candidate’s webcam to assess the self-confidence, attention, reaction speed and stress level of a candidate.
AI components include face recognition, which captures facial expressions, gestures, tone of voice and the frequency of a smile and matches them with the seven universal emotions. Games and videos of typical work situations complete the interview. Altogether the process is supposed to determine how confident, positive and stress-resistant the applicant is.
The whole thing is made possible by so-called affective or emotional analysis: the evaluation of facial expressions, gestures, speech rate, tonality, choice of words and many other characteristics form a profile. How exactly the data points are evaluated by the software is a trade secret(4) and has been called “pseudo-scientific” by opponents.
However, the evaluation of such psychometric tests is firmly anchored in personality psychology and well documented. Of course, HR managers can watch all recorded interviews to make up their own mind.
By the way: With an exaggerated smile, which would appear grotesque in human dialogue, women/men really score points with the software. Restrained micro-expressions are still often too subtle for an AI. That’s why forums and YouTube are now teeming with tips on “correct” behaviour in robot interviews, and universities in the USA offer special training courses.
Brave New World? HR Teams Have to Take a Closer Look Now
Why should HR departments and recruitment agencies have applicants interviewed and evaluated by soulless AI software? Speed, more time for personal contact and a better Candidate Experience. In times of a pandemic, it might be the only way.
Large companies need to attract the best talents by spotting them in thousands of applications — those who smartly automate their application process from pre-selection to the first interview are many times faster than others and enjoy clear competitive advantages in the war for talents. The best candidates don’t want to wait weeks for feedback just because HR has to sift through piles of CVs and can hardly keep up with interview appointments.
Candidates also benefit from this new way of recruiting: they save travel time and costs and have the freedom to conduct the interview at a time that fits into their daily routine — without having to take a leave of absence. In times of curfews and remote work policies, it is the only way to get a job. 80% of all live interviews conducted at HireVue took place outside normal working hours.
Artificial Intelligence For More Diversity — Or Reinforcing Bias?
AI algorithms are also supposed to help in automated personnel pre-selection to assess job applicants without bias and thus ensure more fairness to avoid discrimination based on gender, origin or skin colour. However, even the most modern artificial intelligence cannot guarantee this per se, since it is only as good as the data with which it has been “fed”.
We are talking about machine learning, and such an AI must first be trained for this specific task. This is done by using data such as the CVs of the top achievers working in your company. If, however, these already have certain characteristics due to an unconscious bias of the personnel managers, an AI trained with this data will not only replicate such characteristics but even reinforce prejudices. It makes things worse.
Another problem is that there is still a gender gap in AI research and development teams worldwide, where only 16% of those team members are female. The reason why we need more women in A.I. is that half of the world’s population is not participating equally in research and development of technologies that will have an impact on all of us. We need more diversity to reflect humanity in these decision processes.
Meeting these challenges requires project managers and consultants who are familiar with the subject matter and who work closely with HR and IT in selecting and implementing AI solutions.
High-Tech Needs High-Touch
Especially when recruiting talent, we will always need the human touch. We need this gut feeling and (human) judgement to know if someone actually fits into the team. And we all know that hard skills are only half the battle. Cultural fit, personality and soft skills like self-confidence and emotional intelligence can (still) be judged better by people than the best AI application. But the tasks that AI applications already master well in the application process are completed a thousand times faster than you could ever do it. Your candidates will thank you for speedy response time and you will have more time for personal relationship building and a refined hiring strategy.
And let’s face it: In times like these, we simply have to try out new ways.
About the author
I advise companies in the field of emerging technologies for learning and collaboration in remote teams and accompany you in implementing friendly AI-driven chatbots or seamless integration of AI solutions for HR and L&D.
Advocate for independent AI, privacy, diversity and women in tech. Member of Women in AI and Women in Tech. I work 90% online with clients all over the world.
Let’s talk: gaby@yourvirtualtrainer.net
Let’s connect: https://www.linkedin.com/in/virtualtrainer
(1) 74% Termination rate in the application process
(2) KATY, the career bot at T-Systems, Germany
(3) Technologisation of the application process (Marketinsider)
(4) The company HireVue may soon have to disclose its algorithms, as these are classified as critical by the EU Commission: EU-Kommission will mehr Transparenz bei Künstlicher Intelligenz (“EU Commission wants more transparency in artificial intelligence”, Handelsblatt 30.1.2020)
Don’t forget to give us your 👏 !
