The disclosure came during Google’s Q3 2024 earnings call on Tuesday.
“We’re also using AI internally to improve our coding processes, which is boosting productivity and efficiency,” Pichai stated during the call.
“Today, more than a quarter of all new code at Google is generated by AI, then reviewed and accepted by engineers. This helps our engineers do more and move faster.”
The widespread adoption of AI-assisted coding gained momentum with GitHub Copilot’s introduction in 2021 and its public release in June 2022. The tool initially relied on OpenAI’s Codex model, designed to both suggest code completions and generate new code from natural language instructions.
Since then, the field has seen rapid expansion, with major tech companies including Anthropic, Meta, Google, OpenAI, and Replit developing increasingly sophisticated solutions.
In a recent development, GitHub announced that developers can now use alternative AI models such as Anthropic’s Claude 3.5 and Google’s Gemini 1.5 Pro within Copilot, shifting away from exclusive reliance on OpenAI’s technology.
Despite the enthusiasm surrounding AI-assisted coding, some experts express concerns about potential risks. A 2023 Stanford University study found that developers using AI coding assistants were more likely to introduce bugs while paradoxically feeling more confident about their code’s security.
