45% of India assessments now flag cheating
HirePro's 'Candidate Integrity in Job Assessments and Interviews 2026' report, based on roughly 1 million assessments and 100,000 interviews a year between FY21 and FY26, finds that assessment cheating climbed to 45% in 2025-26 from 32% in 2021-22 — nearly one in two assessments now carries an integrity concern. Interview malpractice rose to 2.5% from 1.4%. The composition is shifting fast: external assistance fell to 43% of assessment cases from 55% in 2022-23, while AI-assisted cheating hit 12% and grew 50% in a year, and video or screen manipulation nearly doubled since 2022-23. In interviews, identity fraud collapsed from 65% of cases in FY23 to 28% in 2025-26, replaced by AI-assisted cheating (18%) and device-based cheating (20%). Key signal: Cheating in assessments rose to 45% in 2025-26 from 32% in 2021-22, with AI-assisted cheating the fastest-growing category at 50% relative growth in a single year. For hiring leaders, this matters because for Indian IT services, GCCs and product firms running high-volume fresher and lateral hiring, the assessment funnel is no longer a reliable filter. When 45% of assessments throw up an integrity concern, the cost lands in three places: wasted panel hours, bad offers, and post-joining performance failures that surface months later. The shift from identity fraud to AI-assisted cheating also breaks the standard control set — ID checks and remote proctoring catch the old fraud, not the new one. TA leaders should assume their current evaluation stack is at least one generation behind candidate tooling, and budget for continuous redesign rather than a one-time proctoring licence. Teksands view: Stop buying proctoring and start redesigning assessments. The fraud has moved from fake IDs to AI assistance inside the test itself, which means the controls most Indian hiring teams bought in 2022-24 are catching yesterday's problem. If your funnel still leans on remote, unsupervised, MCQ-heavy screening for volume tech roles, assume a meaningful share of your shortlist is not who the score says they are. The fix is expensive and inconvenient: synchronous problem-solving, live code walkthroughs, and interview panels that probe reasoning rather than recall. Cheaper than a bad hire at 6-12 months of salary.
Key fact
Cheating in assessments rose to 45% in 2025-26 from 32% in 2021-22, with AI-assisted cheating the fastest-growing category at 50% relative growth in a single year.
Why it matters
For Indian IT services, GCCs and product firms running high-volume fresher and lateral hiring, the assessment funnel is no longer a reliable filter. When 45% of assessments throw up an integrity concern, the cost lands in three places: wasted panel hours, bad offers, and post-joining performance failures that surface months later. The shift from identity fraud to AI-assisted cheating also breaks the standard control set — ID checks and remote proctoring catch the old fraud, not the new one. TA leaders should assume their current evaluation stack is at least one generation behind candidate tooling, and budget for continuous redesign rather than a one-time proctoring licence.
The Teksands point of view
Stop buying proctoring and start redesigning assessments. The fraud has moved from fake IDs to AI assistance inside the test itself, which means the controls most Indian hiring teams bought in 2022-24 are catching yesterday's problem. If your funnel still leans on remote, unsupervised, MCQ-heavy screening for volume tech roles, assume a meaningful share of your shortlist is not who the score says they are. The fix is expensive and inconvenient: synchronous problem-solving, live code walkthroughs, and interview panels that probe reasoning rather than recall. Cheaper than a bad hire at 6-12 months of salary.
Got a tech role that's refusing to close?
Send us the JD. We'll tell you whether the problem is talent supply, compensation, location, process - or the JD itself.
