The AI cheating problem: Google puts an end to virtual coding interviews

Mumbai: Google is reintroducing in-person coding interviews to combat the rising issue of AI-assisted cheating in virtual technical assessments, CEO Sundar Pichai confirmed. With candidates increasingly using AI tools off-camera to answer coding questions, virtual interviews have become unreliable for evaluating real skills. As a result, Google will ensure that at least one round of interviews is conducted face-to-face to better assess a candidate's fundamental knowledge and abilities.

This decision follows internal pressure from Google employees, who have voiced concerns about the integrity of remote hiring during town hall meetings. The company acknowledges that while virtual interviews are faster and more convenient, they are vulnerable to AI exploitation, with reports suggesting that over half of some virtual technical candidates may be using unauthorised AI assistance.

Google is not alone in addressing this challenge; other tech firms like Amazon, Cisco, McKinsey, Anthropic, and Deloitte have also implemented stronger anti-cheating measures. These include bringing back on-site interviews, requiring candidates to sign agreements not to use AI tools, and outright bans on AI use during hiring.

Pichai emphasised that these changes aim to maintain high hiring standards and ensure employees possess strong computer science fundamentals even in an era dominated by AI.