AI-based proctoring versus human proctoring

Proctoring: AI-based proctoring vs Human Proctor

The witty remark that Covid-19 pioneered a vast portion of digital transformation in many business enterprises carries a lot of truth in reality. The inevitable lockdowns and work from homes have fast tracked many technological innovations across multiple industries and domains.

Recruitment industry too has taken on drastic adaptations to suit the ‘new normal.’ It is not only remote sourcing and remote assessment of candidates, but remote proctoring is also gaining momentum. Unfortunately, the rise of remote ‘everything’ has also significantly decreased the seriousness level among candidates. For example, between the end of 2020 to the first half of 2021, Glider Intelligence tracked a 92 percent increase in fraudulent test tracking activities. Further, the AI-enabled proctoring features also reported 23 percent of the candidates engaged in some assessment malpractice.

A Washington Post article cited that cheating has almost increased 8 times during the pandemic.

The traditional standard system of human proctoring is no longer the only invigilation mechanism today. Backed by AI, proctoring systems can be classified into three types:

  • Live proctoring: Real-time proctoring method that requires human proctor to monitor candidates and track their behavior from start to finish. It also facilitates two-way chat between proctor and candidates.
  • Recorded proctoring: Video recording of the candidate assessment. Allows flexibility, like candidates can take exams anywhere, anytime. Also requires human intervention for review.
  • Automated proctoring: The algorithm-powered system is enabled to detect fraud and mischievous activities committed by candidates though audio and video parameters.  Requires no human intervention.

As per an article from The Wall Street Journal, companies who offered software to combat cheating in remote assessments found their revenues grow in the pandemic time.

On the same line, Newyorker.com also reported that companies that designed online proctoring apps saw a surge of 500 per cent in their business over the past year.

AI-based Proctoring

Online proctoring system was in use before the pandemic. Some of the reputed recruiters and universities used online proctoring for their assessments and examinations. But the evident lockdowns and remote working facility brought online proctoring into mainstream.

Artificial Intelligence Proctoring System or AIPS in short has reinvented proctoring mechanism. Apart from being user-friendly, AIPS keeps track of the candidate exceptionally, with a little or no human intervention.

At the outset, AI-based proctoring is an automated virtual proctoring system that monitors the candidate and his environment during the assessment duration. The system mainly deters and detects any suspicious candidate activity that amounts to cheating.

The Concept Map of AIPs below depicts how it works.

Source: A Systematic Review on AI-based Proctoring Systems: Past, Present and Future

DOI: 10.1007/s10639-021-10597-x

Features of AI-based Proctoring:

  • Two-factor authentication
  • AI facial recognition
  • 360-degree environment surveillance
  • Real-time monitoring
  • Browser lockdown
  • Network traffic restriction
  • Webcam and microphone analysis
  • Multi device compatibility
  • Auto-flag suspicious activities
  • Assess any false-positives
  • Test integrity score
  • Detailed report generation
  • GDPR compliance

How does AI-based Proctoring work?

Compared to other proctoring systems, AIPS requires very less human interference. It relies more on technology. Certain specific requirements of AI-based proctoring are:

  • Uninterrupted web connection
  • Plug and play technology
  • Web camera
  • Microphone
  • Screen share
  • Question bank
  • Hardware add-ons (like speech recognition and gaze tracking)
  • Trained proctors (for human oversight)

Human Proctor

The classic invigilation system. We have all been a part of this from schools to universities. Just because the world has transitioned to AI-based automated proctoring, it doesn’t mean human proctoring is obsolete. While it’s true that much of the tedious activities are taken care by technology, AI still needs the watchful eyes of human proctor.

A human proctor is an individual who invigilates candidates on real-time. They monitor a candidate or multiple candidates from start to finish. Human proctors need to be professionally trained with online proctoring technologies. For this reason, many of the recruitment companies hire third-party professionals who are well-versed in the job.

In short, human proctoring is a live proctoring mechanism who monitor the candidates online unlike the physical test centers.

Importance of Human proctor in AIPS:

All kinds of proctoring systems including the AI-based are human centered. According to Scott McFarland, CEO of ProctorU, “AI can never be a complete substitute for human proctoring. A trained proctor is necessary to correct behavior, review and make a judgement call on.”

There are many view points on how much AI and human proctoring should be integrated and if it is legal and ethical to allow AI proctor on its own completely. Nevertheless, AI or any technology is meant to support human decision making, but not entirely substitute the latter.

Similarly, it is never a debate to discuss which proctoring is better: AI or human. Human intervention is necessary for candidate authentication and no technology or algorithm does a better job of finding talent quality than human judgments and decisions.

There are many examples of how AI and ML went wrong when they acted completely by themselves. The infamous case of how Amazon’s hiring algorithm failed to be gender-neutral is an eye-opener for the proctoring domain too. Although with no fault at the company side, (the algorithm simply mirrored the past job applications from female candidates – which was low).

A human proctor is significant because of the following reasons:

  • Address technical issues and errors
  • Chat in case of dilemma and confusions
  • Friendly conversational tone
  • Calm candidate’s anxiety and distress
  • Maintain brand reputation
  • Clear false positives
  • Analyze the data for misconduct or integrity violation
  • Final review and decision

Before We Close

True to its belief, “Talent First and Integrity Always”, Glider allows technology and human intervention to combinedly empower proctoring mechanisms. In one of the SIA’s interview, Glider CEO, Satish Kumar reiterated that contrary to popular myth that AI perpetuates unconscious bias in hiring decisions, Glider AI does not let the system decide. It only troubleshoots the frauds and deceits. The recruiters are presented with the findings. In this regard, the ultimate decision of talent recruiting lies in the hands of the hiring team always.

Glider’s proctoring tools like automated monitoring, comprehensive monitoring report, and auto flag cheating mechanisms are intensely data-driven. The recruiters can review the report and take enhanced and sound decisions on the candidates. This is how Glider juxtaposes AI and human proctoring to makes blended decisions on candidate authentication and talent quality.

Leave a Comment

Your email address will not be published. Required fields are marked *