The State of AI Interview Cheating in India 2026.
A 12-page research report based on 200+ AI voice interviews we ran in Q1 ’26 across Indian IT, BPO, and engineering teams. What we found about how candidates cheat — and the 7 signals our customers use to catch them.
14% of senior-engineer interviews use ChatGPT mid-call
Across 200+ interviews we ran in Q1 ’26, response-timing patterns + vocabulary shifts caught hidden AI assistance in roughly 1 in 7 senior-IC conversations. Most teams catch zero today.
Candidates cheat least when interviewed in their own language
Malayalam, Hindi, and Tamil interviews showed 4× lower fraud rates than English-only screens for the same role — when candidates aren’t code-switching, they don’t need ChatGPT.
BPO + customer-support roles show the highest cheating rates
23% of BPO interviews flagged at least one fraud signal — likely because the interview format (script-heavy, English-only) is most vulnerable to AI ghostwriting.
One-time face verification catches almost nothing
Of 14 confirmed cheating cases in our dataset, 13 had passed initial face verification cleanly. Continuous monitoring caught all 14. The proctoring you're paying for probably isn't working.
Phone-screen replacement saves recruiters ~₹3-5L per quarter
Median Kerala IT team running 30 hires/quarter saves ~250 recruiter-hours by replacing phone screens with AI voice interviews — about ₹3-5L of recruiter cost depending on team size.