Nearly 50% candidates cheat in online assessments. Here’s how AI can help
Mankind has entered the era of Artificial Intelligence. While it’s not perfect yet, it can still resolve several issues effectively in no time. In the midst of the war for talent, not only it’s hard to find the right people but also guarantee if he or she is the right fit. An analysis of assessments conducted by HirePro (an automated hiring platform) revealed that 30-50% of people cheat at entry-level job assessments.
Some other findings are that cheating detection is most efficient when the proctoring combines all three forms - video, audio, and image. This combination is highly effective for online assessments as only 2% of cheating gets missed out. When the proctoring combination is of only audio and image, only 58% of cheating is not detected. And with image-only proctoring, almost 92% of cheating goes undetected.
To understand more and deliberate on fraudulent practices by candidates during job assessments in India, People Matters exclusively spoke to Mr S Pasupathi, the Chief Operating Officer at HirePro, an automated hiring platform.
Excerpts from the interview:
How has AI automation changed the way tests and interviews are conducted?
Before the pandemic, assessments and interviews were conducted physically, in the presence of a proctor, or by an interviewer. For lateral hiring, companies used to call candidates to their office premises for the interview, and for campus recruitment, organisations visited the college to conduct the entire hiring process.
But with the help of AI and automation, tests are held online and are organised remotely. Both organisations and candidates prefer online assessments, and scheduling in-person interviews continue to be challenging. In recent times, organisations have stopped visiting colleges for campus hiring as they find virtual hiring cost-efficient and time-saving.
The usage of automation and artificial intelligence (AI) based proctoring technologies helps prevent malpractice in online job assessments. The tools are designed to deliver secure and reliable results by weeding out instances of malpractice. Companies also leverage a combination of AI and manual review to detect fraudulent practices and ensure the integrity of online assessments. Organisations rely on AI-based interview tools and platforms for fraud-proof interviews.
Why do you believe AI and automation are not enough to judge malpractice?
Candidates who wish to cheat will always try to find ways to do so, be it the offline or online mode of tests. The effectiveness of the proctoring solution is also a function of the infrastructure readiness and solution chosen by the organisation. A lot of companies currently use basic proctoring like taking photos of the candidate rather than a full video and audio.
This gives scope for the candidate to cheat. Ideally, if the candidate is under the surveillance of two cameras then the best results can be obtained. Live Proctoring solutions are also available in case it is a high-stakes examination. AI is designed to assist human beings in making better decisions, and should not by itself be allowed to make decisions.
Instead, it can be used to reduce human effort, improve accuracy, and flag instances of malpractice. With the evolution in technology and the advent of AI, the ability to detect and measure cheating during online assessments has increased. Organisations can use a combination of AI and manual review to detect fraudulent practices.
However, in many real-life problems, there are a good number of corner cases that automation and AI will never be able to solve.
According to a HirePro report, 30-50% of people cheat at entry-level job assessments. Who is to be blamed for this and what can be done to improve this?
We cannot prevent candidates from cheating completely in online job assessments as they find ways to get ahead. Candidates tend to figure out their luck to cheat and pass the test. Candidates with lower skill proficiency and those who are not updated with the current technology or have low expectations of success due to a perceived lack of ability are more likely to cheat. Employing automation and AI helps organisations prevent malpractice in online job assessments.
What are some of the tools to detect malpractices?
There are various tools and methods to make online job assessments cheat-proof. Firstly, we ask candidates to use a secure browser and secondly, we use proctoring to detect cheating. In the third method, a physical proctor monitors candidates' movements when they're taking the assessments. AI-enabled talent assessment tools help organisations to detect fraud and malpractices in the hiring process.
Without human intervention, are we compromising a quality talent pool and unknowingly discarding potentially good candidates?
AI and automation help us to objectively assess a candidate. For example, when a person writes code, we can figure out their technical expertise. Here, AI helps us to figure out any kind of malpractice that's happened. The right set of candidates without fraudulent activity will be able to succeed based on their assessments. And additional human intervention, at the end of the day, is finally deciding the competency of these candidates.