Corporate India closing interview doors to prying AIs

Bengaluru: Companies in India are increasingly deploying countermeasures to screen out job seekers misusing artificial intelligence (AI) tools to game virtual interviews. Recognising that many candidates now indulge in AI-assisted cheating, companies such as Deloitte, Deutsche Bank, Scaler and Meesho are keeping an eye out for highly polished responses that lack depth, delayed answers or unnatural conversational patterns and difficulty in explaining logic or assumptions, according to industry executives. They are also putting in place a slew of measures to assess candidates' genuine problem-solving abilities and implementing advanced monitoring capabilities to prevent interview malpractices. Deloitte is now using more scenario-based and case-led discussions, where candidates must explain approach and assumptions. Probing questions and follow-up discussions assess how they think and respond under ambiguity, areas where genuine capability becomes more visible, said Deepti Sagar, chief people and experience officer, Deloitte India. “Multi-stage evaluations with cross-functional interviewers help corroborate our observations,” she said.

An analysis of more than 20,000 real interview records by interview intelligence and recruitment automation platform InCruiter indicated that nearly one in three candidates directly leveraged AI tools during live interviews. “Interview malpractice is no longer limited to proxy candidates or external assistance. We are witnessing candidates using hidden browser tabs, second monitors with ChatGPT running, and voice cloning software to bypass evaluations," said Anil Agarwal, cofounder, InCruiter.

At Deutsche Bank, the approach to countering AI-assisted cheating includes assessment platforms with built-in safeguards that can identify unusual response patterns and flag potential inconsistencies and competency-based interviews that go beyond theoretical knowledge. “The assessment platforms we use are equipped with advanced monitoring capabilities, and are designed to automatically detect and flag anomalies like the use of unauthorised browser extensions or unusual activity in real-time,” said Madhavi Lall, head of human resources, Deutsche Bank Group India.

Hiring processes at ecommerce company Meesho are evolving through AI-powered assessments with human-led evaluations. The company uses AI-assisted interview frameworks, secure browser environments for virtual interviews and nd advanced video analytics to enhance assessment quality and integrity.

“These help us evaluate candidates more holistically while also identifying inconsistencies, unusual patterns or potential misuse of external tools during assessments. We are also designing interviews around real-world problem solving, case-based discussions and collaborative exercises that reflect how work gets done today,” said Ashish Kumar Singh, chief human resources officer at Meesho. Recruiters, however, do not see AIassisted cheating as a big setback. It is unlikely to reverse the overall shift toward flexible and digital hiring models, with remote hiring remaining essential to accessing diverse talent pools, said Deloitte’s Sagar.

Edtech platform Scaler is redesigning assessments to evaluate higher-order thinking, judgment, problem-solving depth and practical application, areas where over-reliance on AI without real understanding manifests quickly. “Like many companies, we have observed occasional instances where candidates appear to use AI tools in real time during interviews or assessments. However, rather than treating this only as a policing problem, we believe the larger shift is that companies will increasingly need to evolve assessment methodologies themselves, since traditional questionanswer formats are naturally more vulnerable to AI-assisted responses,” said Akshay Raje, head of people and culture, Scaler. As a result, companies’ hiring processes will move toward more contextual discussions, live problem-solving, scenario-based assessments, practical simulations, collaborative exercises and deeper evaluation of thought process rather than just “correct answers,” he added.