Google says 75% of internal code is AI-generated

Google’s AI-assisted coding surge inside the company

Google says that 75% of new code created inside the company is now generated by AI and reviewed by human engineers, up from 50% last fall. The message is that AI coding tools are no longer limited to experiments or prototypes; they’re becoming part of everyday software development workflows, with humans continuing to provide the final review.

What changed

• Teams increasingly rely on AI to draft code faster.
• Engineers still review that output, keeping human oversight in the loop.
• The shift from 50% to 75% suggests adoption accelerated over the past year rather than trickling in.

Why it matters

This matters because software development is a major cost center for tech companies, and the ability to increase throughput can influence everything from product release cadence to engineering hiring demand. It also signals a broader industry pattern: “AI coding” is evolving from stand-alone tools into integrated processes that shape how engineers plan, write, and validate code.

At the same time, the headline implicitly highlights a governance tension. Even if code is AI-generated, the work doesn’t disappear—review and quality control remain central, which is why Google’s phrasing stresses that engineers are still checking what the models produce.

Overall, the report positions AI coding as a mainstream productivity lever inside a major vendor’s engineering organization, not just a consumer-facing feature or a demo for investors.