A team of NC State NSF Research Assistants delivered a captivating panel presentation about AI-Assisted Recruitment Screening Tools, Processes, and Strategies (slides) at the 2020 Symposium on Communicating Complex Information (SCCI). The 9th annual conference was held at Old Dominion University on February 24-25, 2020.
The NC State NSF Workforce Development panel included Principal Investigator (PI) and organizer, Dr. Huiling Ding, Ph.D. students Yeqing Kong and Jet Wang, and Master students Hunter Jones, Kelia Ray, and Chenxing Xie. Professor Ding and Kelia Ray joined the presentation remotely from the NC State English Department, in Raleigh, NC.
The Symposium on Communicating Complex Information (SCCI) explores how the design and presentation of complex information affect how humans use it. Information complexity arises from the inherent relationships between context, content, and people using the information.
The panelists discussed various AI-assisted recruitment tools and strategies to deal with new challenges.
- Automated Video Interviews
It has been a while since organizations used pre-recorded questions for video interviews to screen job seekers, while some recent works have been done to even automate the screening process. The algorithm will analyze both the content and non-content information and can thus generate the hireability scores (Muralidhar, Nguyen, Gatica-Perez, 2018). For content analysis, verbal answers will be transcribed into texts, while facial expressions and speech qualities might be used to infer personality or interpersonal skills for non-content analysis (Chen et al., 2016; Muralidhar et.al., 2018; Naim, Tanveer, Gildea, & Hoque, 2018). Despite the convenience along with the automation, it might put exceeding weight on job-irrelevant traits or skills and thus lead to potential adverse impact and fairness issues (Blacksmith, Willford, & Behrend, 2016; Hiemstra, Derous, Nikolaou, & Oostrom, 2015). Therefore, professional communicators can assist job seekers with question comprehension, answer organization, and improvement in non-verbal performance (e.g., speech fluency, emotion expression, and body movement).
- LinkedIn Optimization in the Era of Artificial Intelligence
As 94% of recruiters use LinkedIn to search for potential candidates (Rangel, 2014), creating an effective LinkedIn profile becomes a critical task for job seekers. While there is an array of strategies for LinkedIn optimization regarding verbal and visual content and Search Engine Optimization (SEO) (Berk, 2013; Cooper, 2014; Thies, 2012), few of them address LinkedIn optimization in terms of AI recruiting. With the deployment of Talent Search system at LinkedIn, the features of this AI recruiting system, including intelligent query understanding, the mutual interest matches between recruiters and candidates, and the personalized preference models of recruiters (Geyik et al., 2018; Ha-Thuc et al., 2015; Ha-Thuc et al., 2016; Ozcaglar et al., 2019), brings both opportunities and challenges to job seekers. Thus, this study proposes novel LinkedIn optimization strategies, such as modeling their profiles after current employees in target companies and using standard words in the profiles, to help candidates address the challenges brought by the tide of Artificial Intelligence.
- Being Strategic to Succeed in Social Profiling
Profiling is both a practice and a technique that involves “automated processing of personal data” aimed at developing predictive knowledge in decision making in various domains such as employment screening (Council of Europe Recommendation, 2010; Ferraris et al., 2013). Despite the ethical and legal issues brought by social profiling (e.g., prejudice and discrimination) (Mitrou et al., 2014), increasing hiring managers are integrating AI tools to predict applicants’ personality and behavior based on social media and other public information (Harwell, 2018; Thibodeaux, 2017). To help the job applicants succeed in AI-assisted social profiling, professional communicators should provide training on managing applicants’ online presence, such as “educat[ing] applicants about strategic placement of personal information” on social networking sites profiles (Evuleocha & Ugbah, 2018).
- Neuroscience Games as an Attractive and Ethical Model for Recruitment
The emerging field of neuroscience games offers an attractive model for AI recruitment purposes. Psychometric testing has already found its place in the world of recruiting. However, evidence suggests that current tests produce false responses (Wintersberger, 2017) and that the questions being asked may not be effective (Cripps, 2017). The introduction of neuroscience games presents an engaging method of recruitment that allows applicants to use games as a way of measuring aptitude and personality rather than a series of questions. Asking questions, just like an interview, presents a “managed interaction” in which applicants put on performance rather than demonstrating their true capabilities (Wintersberger, 2017). Gamifying the recruitment process is backed by data that suggests that this produces a proactive rather than reactive applicant and higher job compatibility rates (Chamorro-Premuzic et al., 2016). These benefits find themselves tested against ethical issues in the field of AI recruiting including privacy, bias, and the resistance of those seeking a more traditional method of job search. An examination of the anti-bias initiatives within programs like pymetrics as well as critical studies of neuroscience games reveals a highly effective system that highlights focused applicants while paving the way in anti-biased recruiting.
The future of recruitment is uncertain with AI tools profoundly changing how job seekers are screened, evaluated, and selected. AI-assisted technology offers benefits that include consistent evaluation processes, better utilization of time management, and the ability to match job seekers with relevant job options. Recruitment tools have been shown to help identify top performers with the use of personalized networking sites and more direct access to job recruiters. Biases such as age, digital gap, gender, and ethnicity must be considered with each AI tool and strategy presented. Consider three notable’ questions.
- How might we use AI-assisted recruitment tools to help redefine the field of workforce recruitment?
- How might we help job seekers benefit from AI-assisted recruitment strategies and tools?
- How might these emerging technologies help existing employees advance from within more effectively?