Google Tells Staff to Get Approval Before Using External AI Coding Tools

In a recent directive, Google has instructed its software engineers to obtain approval before using external AI coding assistants. The guidelines come as the tech giant encourages employees to prioritize its internal AI tools, according to reports.

Google Engineering Vice President Megan Kacholia issued the directive in a June email, urging staff to rely solely on internal AI models for coding tasks. Engineers wishing to use third-party AI tools for non-coding purposes must also secure management approval, the reports stated.

The policy reflects Google’s broader strategy to enhance productivity through AI while maintaining strict control over sensitive internal information. Reinforcing this message during an all-hands meeting in July, CEO Sundar Pichai emphasized that employees should use AI tools to help Google compete with rivals who are also leveraging artificial intelligence.

Google has been actively promoting its AI coding platform, Cider, which runs proprietary models including “Gemini for Google” – formerly known as Goose – trained on the company’s internal technical data. Since its May launch, half of Cider’s users access the service weekly.

As Google competes with firms like OpenAI in AI coding, it has invested $2.4 billion to acquire key talent from AI startup Windsurf, including CEO Varun Mohan, to advance its “agentic coding” capabilities. Engineering managers have asked employees to demonstrate daily AI usage, though reports indicate that AI adoption is not formally evaluated in performance reviews.

Pichai noted that over 30% of Google’s code is now AI-generated, up from 25% last October, contributing to an estimated 10% productivity boost among engineers. Like Amazon and Microsoft, Google is encouraging widespread AI adoption to maintain a competitive edge in the rapidly evolving tech landscape.