"Human-Led, AI-Assisted"... Google Allows AI Assistance in Developer Interviews

Google Introduces AI Tools in SW Engineering Interviews
Comprehensive Evaluation of Prompt Engineering and Debugging Skills
Meta Also Piloting "AI-Assisted" Coding Interviews

Google has decided to allow candidates to use artificial intelligence (AI) tools during software (SW) engineering job interviews.

Google logo image. Photo by AP Yonhap News

According to Yonhap News on May 7 (local time), citing U.S. business media outlet Business Insider, "Google is conducting a pilot program for a new interview process that allows SW engineering candidates to utilize AI tools."

Under this process, candidates can use AI tools in the 'code comprehension' assessment to read, debug (fix errors), and optimize existing code databases. Interviewers evaluate the candidates' ability to craft effective prompts for AI, as well as the quality of their output and their debugging skills. The AI tool candidates will use is expected to be Gemini, Google's proprietary model. Google plans to first implement this system for entry-level and mid-level positions in select teams in the U.S., and then expand it company-wide depending on the results.

Google explained that this measure reflects the current workflow of the era, which is "human-led, AI-assisted." Brian Ong, Vice President of Talent Acquisition at Google, stated, "We have introduced a pilot program for SW engineering interviews to match the AI era," adding, "We are constantly improving our interview processes to identify the best talent."

The reason Google is incorporating AI into its hiring process is because the way SW engineers work has been rapidly changing. Last month, Google revealed that AI now writes three-quarters of the new code within the company.

Photo of Google's generative AI Gemini to aid understanding of the article. Photo by Reuters and Yonhap News Agency

Graphic design startup Canva is already operating a similar hiring process. Since last year, Canva has introduced an "AI-assisted coding round" in its engineering interviews. The company has eliminated traditional algorithm memorization tests and now has candidates use their preferred AI model to complete real product tasks. Canva explained this change by saying, "Since our engineers already use AI to analyze large code bases and generate code, interviews should reflect actual work."

AI coding startup Cognition also operates a hiring process that allows the use of AI. Emily Cohen, Head of People and Operations at Cognition, compared requiring candidates to take coding tests without AI assistance to "taking a math test without a calculator." She emphasized, "Candidates should be able to use—and must use—AI tools to perform tasks similar to what they would do in the actual job."

Meta has also reportedly been running a pilot program with a similar approach since October last year. In this process, one of the two traditional coding interviews is replaced with a test where candidates spend 60 minutes in a dedicated coding environment equipped with an AI assistant. The available AI models are said to include OpenAI's ChatGPT series, Anthropic's Claude, and Google's Gemini.

This content was produced with the assistance of AI translation services.