Busting myths about online proctoring
Last year, the outbreak of the Covid-19 pandemic profoundly altered the way business is conducted. Talent acquisition is no different. Online proctoring, for instance, was practised even during the pre-pandemic times, but its adoption for conducting multiple examinations and tests virtually has increased multifold amidst the pandemic. Since online proctoring solutions and technologies are constantly evolving, there are certain common misconceptions about the technology and the process. Business leaders and recruiters need to understand them as they look to adopt automated online proctoring.
Online proctoring alone cannot eliminate fraud
Some candidates can manage to indulge in malpractices even under the strictest of physically-proctored assessments. Similarly, it can happen during online proctoring too. While it is fair to expect online proctoring to drastically reduce the number of fraudulent cases, it cannot eliminate them completely.
- It is unfair to rely on artificial intelligence (AI) to auto-reject candidates who display suspicious behaviour during or after the test/interview for the following reasons:
- Global AI regulations explicitly state that a negative decision cannot be taken by AI without a human review.
- Auto-rejecting the candidates based on AI-detected anomalies can hurt genuine candidates.
- AI cannot make a perfect decision in all conditions for all candidates, and hence, some false positives are to be expected.
Eliminate unintended AI bias
It is not wise to rely on AI alone for auto-detection of suspicious eye movements, facial expressions, and body language. A genuine candidate and a fraudulent candidate could have the same eye movements, facial expressions, and body language.
Besides significantly increasing the manual review effort of the recruiters, flagging candidates as suspicious based on these anomalies can put them at the risk of unintended AI bias. People of different genders, ethnicities and regions can have naturally different eye movements, facial expressions, and body language.
Prudent proctoring and privacy
- Expecting AI to auto-detect whether the candidate is using a mobile phone or a book to refer to
- Answers are best avoided. A genuine candidate may have a genuine need to use a mobile phone for the internet (mobile hotspot), or a notebook/scratch-pad to answer the questions. The shape and appearance of mobile phones and books are quite similar to other everyday objects like a tissue box, newspaper, photo frames, and small furniture items.
- Marking candidates as suspicious based on these anomalies will significantly increase the manual review effort of the recruiters. From a privacy and compliance standpoint, organisations will need explicit consent from the candidates to be able to analyse their surroundings and the objects.
The way ahead
As the pandemic is resetting business processes, leaders must reinvent themselves and rethink the workforce’s established practices. Talent-acquisition teams should be equipped with the necessary resources and tools to navigate the new workforce management trends and drive the organisations’ success.
Online proctoring has a critical role to play. It creates a level-playing field for candidates, recruiters, and test administrators. It allows everyone to be judged on their merit and helps recruiters and interviewers focus their efforts on genuine candidates. It helps businesses enhance their revenues and brand equity. It is all achievable, thanks to a combination of AI technology, human intervention, process design, and appropriate expectations.