DS Cohen & Associates
While most of the world remains working virtually, the ritual of hiring and onboarding new talent continues. The disruption of the pandemic to how many things get accomplished has not disrupted the basic need to hire new employees. The last 20 months have accelerated a significant number of new tools and technology designed to enhance the hiring process.
The headlines read, “Companies turn to AI hiring tools to manage the volume of applicants and remove human bias.” One of the most significant of the tools is the application of Artificial Intelligence (AI). One of the LinkedIn Learning interview programs claims that the use of virtual video interviews significantly reduces bias. But is this a panacea or just another snake oil solution for the issue of hiring without bias?
Can AI Decode Your Personality?
Another of the tools is robotic interviews combined with word analysis of the responses letting you know that the person fits your requirements or not. Companies with significant hiring needs are embracing these tools. Among those is an exciting tool from Nemesysco, a leader in voice analysis solutions that focuses on Layered Voice Analysis (LVA). The technology goes beyond the visual and verbal expression and focuses on tonality to provide insight into the candidate’s personality trades. The tool goes beyond the assessments relying on self-reporting from the candidate, the same person you’re trying to assess. To summarize the technology, it analyzes the candidate’s tone of speech and correlated vocal data with human emotions for assessments and personality tests.
The LVA W picks up the minute elements from the voice, and the technology is like using a voice microscope. Using tonality, the AI can read your traits, even if you realize you need to change your focus or story. The AI can pick up 151 biomarkers that we pick up from the voice. The tool focuses on emotions such as excitement, stress, confusion, uncertainty, anticipation, hesitation, concentration, mental effort, imagination, anger, and happiness. In simple words: Layered Voice Analysis (LVA) technology correlates vocal data with human emotions for use in assessments and personality tests. (Warning this tool is not to be used in North America for recruitment and selection.)
Today’s marketplace is beginning to flood with tools focusing on interpreting the tone of voice to body language to word analysis. For example, the Educational Testing Service (ETS), the people who write the Scholastic Aptitude Test (SAT), are developing an automated scoring of interview videos using Doc2Vec Multimodal Feature Extraction Paradigm.
The ETS approach uses video-based job interviews that automatically evaluate interview performance. The ETS AI draws its data from monologue-style responses to structured interview questions; the system uses the candidate’s words as the foundation of the report. The report assesses the candidate’s responses for the knowledge and skills and provides holistic judgments of person-job fit. The aspiration is to create a standardized video interview protocol and human rating rubrics focusing on verbal content, personality, and holistic judgment.
The foundation of the ETS program IS a novel extraction method using “visual words” automatically learned from video analysis outputs and the Doc2Vec paradigm. Based on promising experimental results, the hope is this novel method will provide effective representations for the automated scoring of the interview conducted via video using an android asking structured interview questions.
I am in the process of research a new book that will update my book, The Talent Edge, to my current thinking of preparing, conducting, and scoring candidates. A new addition is the use of video conferencing for conducting interviews and applying various new tools. As technology companies flood the recruitment and talent industry, sending mass emails on their tools to make the hiring process more organized and less biased, I thought it would be good to share some of my findings and concerns.
The following is a list of potential issues:
1. To hire right the first time, you must ensure you ask structured behavioural interview questions about your company’s values/culture. The need to hire for fit to culture is a business activity. It is not about hiring in one’s own image and not an obstacle to inclusion and diversity. Instead, it ensures the culture is strengthened, and people trust and respect one another. It makes the process more fact-based and less objective.
2. Hire for fit to the culture: Because of the need to hire for fit to the culture, you need to know the behaviours that define the values that underpin the culture. Since no two organizations have the exact same behaviours, you will need to program each AI tool to the specifics of your company’s culture. Adding the behaviours unique to your company means specific programming. It also means that the person doing the programming must understand how your employees live the behaviours. Speaking with some of the firms developing these tools, they feel this will be difficult and cost-prohibitive. So, like the off-the-shelf psychometric test, you have to go with what the testing company tells you is the right candidate.
3. Voice recognition technology: The issue of accent comes into play. Translation and voice recognition technology are advancing every day. Will an individual mispronunciation of an English word or accents end up correctly translated? Have you ever entered a phrase or sentence into Google Translate? When you look at the results, knowing the language to which it is translated, are the results how you would translate the words into English? I find this problem because we see a translation as nothing more than the person’s perspective doing the translation.
4. Using resume readers: Is using resume readers to eliminate human subjectivity from the employment screening process beneficial? Consider the Amazon experience. When utilizing an applicant scanning process, the realization was, they are seeing far more males than females for interviews. Upon investigation, the programming was the cause. Why? Because without even realizing they have a bias, people bring to work their prejudices; Amazon had to abandon a significant investment. Will this and other applications hinder inclusivity efforts? Remember, the programs are only as equitable as the data that they input. Is the system you purchased perpetuating prejudice and compounding a lack of diversity in your company?
5. Using the Internet: In the effort to save money and recruit from a global applicant pool, and accelerated by higher-speed Internet, the virtual interview has become a norm. In theory, the more advanced Internet quality should result in higher audiovisual quality. It should also provide a higher quality of the visual picture. The Internet connection of one or the other side is sometimes faulty—resulting in interruptions and stoppage mid-sentence. The research concluded hiring managers are considering candidates negatively if the video connection is uneven and choppy; despite being explicitly told to disregard AV quality experiences. In short, video interviews favour job candidates with better Internet connections, and that being aware of this bias does not make it go away.
6. Connectivity of Different Technologies: Being an Apple Computer enthusiast, I am used to the frustration of connecting with other technologies. While this compatibility issue has almost disappeared, it has not for Internet interviews. Some of the applications do not work on specific Internet browsers. Others might initially work but cause issues. Sometimes, the candidate might not have an application being used for the interview and must upload it and become familiar with the technology, unfortunately, during the interview. When you take the issue of #5 above with the internet speed and reliability issue, how might these things impact the ability of AI to ‘translate’ accurately the tonality, emotions, body language, and even the facial expressions of the candidate? Hiring managers do not seem enthusiastic about scheduling a second interview that begins from the start. Regardless of the quality of the application, audiovisual fluency remains an issue.
7. The employee experience: What is the employee experience you are trying to create? What will be the visceral response of the candidate when an Embodied Virtual Agent, i.e., an android, conducts the interview? Versus the experience of being interviewed by a human live and virtually in person? What is the employer brand image you wish to project? Is the image you are projecting consistent with the values/culture of the company?
8. Facial recognition: With the explosion of facial recognition, can the interview tools using facial recognition give you an accurate assessment of the emotional state of the candidate? Facial recognition technologies have inherent flaws. Identification errors are producing false positives and negatives, particularly among minorities. Considered alongside bias in the initial application screening, this presents a compounded threat to organizational inclusivity efforts.
There are many more issues that I will address in the book. The consideration for companies, at this point, is the ethical implications of using a technology you cannot control.
While these advanced tools can undoubtedly expedite the hiring and recruiting process, organizations must consider long-term consequences. In recent weeks, organizations have taken a closer look at standard operating procedures and with diversity and inclusivity in mind. At times, this might require organizations to scrutinize more than their basic hiring practices.
Improving the rigour in the talent process is the first step in improving diversity and inclusion. Technology cannot solve a broken culture or a broken hiring process. Leaders and HR professionals need to commit to a fair and equitable talent lifecycle based on the company’s talent philosophy. The people involved in selecting and using the technology must understand its value and continuously monitor improvement opportunities.
Can AI be used equitably in the hiring process?
With the heightened sensitivity of corporations to diversity and inclusion, these tools are undoubtedly attractive. They say they will eliminate bias and ensure more of a transparent selection of candidates. First, the ethical questions need addressing, or the company might face a longer-term negative consequence. Can your company introduce AI hiring tools fairly and equitably? Can you provide the candidates with the time and resources before the interview to ensure the actual real-time hiring manager and candidate exchange is smooth and technically sound?
Should companies rush into technology applications because they are the “in thing” or make you look up-to-date and cool? Will the flashing lights and slick visual programming offer you magic machines that solve your selection problems or only speed them up? Should the company first think to evaluate the ethical implications of the applications, the short term and longer-term issues and only afterwards set in a policy on using any application? Regardless, after you make a hire, what remains is knowing if you made a quality hire. The measurement of the quality of the employee will remain a lagging indicator.
Even if the technology is more objective and a faster means of decision making, can the tool be designed to meet the specific behavioural critical of the culture of each particular company? How much extra will the customization of the AI be to accomplish this cost? Will using these AI tools mean companies in different industries, even the same industry, make hiring decisions using the same criteria?
You want to ensure that all people have equal opportunities for a job with your company and that neither your recruiters, hiring managers and technology are perpetuating unseen biases.
Keeping the Human in Human Resources
Next week I will provide you with suggestions of steps you might implement to ensure that technology and virtual interviews go smoothly and possibly remove all biases. Remember the interview is still the foundation for the final selection decision. Without providing your hiring panels with meaningful live interview training, without defining, in advance of the interview, the behaviours that represent the values, and without having a scoring system based on the frequency of behaviours, you will not end up hiring the candidate with the best fit to your culture.
___________________
Contact me if you wish to chat. Those who know me know I’m open to sharing and asking tough questions. I always learn so much from the questions people ask and the exchange that follows. Please read my book Inside the Box for a more in-depth read on values, culture and leadership as a competitive advantage.
