Sundar Pichai says 75% of new Google code is now created by AI, marking a major shift in how software is built

Sundar Pichai has revealed that Artificial Intelligence is now playing a central role inside Google, with 75 percent of all new code at the company being generated by AI systems and later reviewed and approved by human engineers. The disclosure offers one of the clearest signs yet of how rapidly AI tools are transforming the daily operations of the world’s biggest technology companies.

The statement was shared in a Google blog post published on April 22 and later highlighted in fresh reports on April 27. According to Pichai, the figure has risen sharply from around 50 percent last autumn, showing how quickly the company has expanded the use of AI in engineering workflows.

This development reflects more than simple automation. It signals a structural change in software creation, where engineers increasingly work alongside intelligent systems that can write, test, refactor and migrate code at scale.

Pichai said Google aims to become “customer zero” for its own technologies. In the business world, that term refers to a company using its own products internally before releasing them widely to customers.

For Google, this means deploying AI across coding, operations, marketing and productivity teams to test real world usefulness before commercial rollout. It also gives the company direct feedback on how its tools perform under large scale enterprise demands.

That strategy matters because Google is competing intensely in the AI race against Microsoft, OpenAI, Meta and Amazon. Internal adoption allows Google to improve products faster while showcasing measurable gains.

One of the most striking examples shared by Pichai involved a complex code migration project handled jointly by AI agents and human engineers. According to Google, the work was completed six times faster than what human teams alone could have achieved just a year earlier.

That claim highlights the next phase of enterprise AI: not merely chatbots or assistants, but autonomous digital task forces capable of handling structured technical work.

Pichai described this as a move toward “agentic workflows,” where multiple AI systems coordinate tasks with human supervision. In practical terms, this could include reviewing code dependencies, rewriting outdated software components, running tests, documenting changes and escalating exceptions to engineers.

For a company operating products used by billions of people, faster internal development can create significant competitive advantages.

Even as AI writes a large share of code, Google says human engineers still review and approve outputs. That point is crucial because software errors at large scale can create security, privacy and reliability risks.

Industry experts have repeatedly warned that AI coding tools can sometimes generate flawed logic, vulnerable code or inaccurate outputs. There have also been cases across the wider industry where autonomous coding tools damaged repositories or removed critical files.

Google executives insist that engineering fundamentals remain unchanged. Richard Seroter said AI is helping developers spend less time on repetitive tasks and more time on architecture, design and solving complex problems.

According to Seroter, software engineers are increasingly evolving into product engineers and architects, focusing on systems thinking rather than routine line by line coding.

The use of AI at Google is not limited to engineering. Pichai also said the company’s marketing teams are using AI tools to generate thousands of creative asset variations at speed.

Processes that once took weeks can now be completed 70 percent faster, according to the figures shared by Google. The company also reported a 20 percent increase in conversions after adopting these AI driven workflows.

That suggests Google sees AI not only as a productivity tool, but also as a direct revenue and performance driver across business functions.

For advertisers and marketers worldwide, the message is clear: creative production at scale is becoming faster, cheaper and more data driven.

Google is not alone in pushing aggressive internal AI adoption. Leaders across the technology sector have spoken openly about integrating AI into everyday work.

Jensen Huang has encouraged employees at Nvidia to use AI whenever possible. Salesforce executives have also discussed rising internal use of automation tools. Reports have indicated that Meta is tying some employee performance metrics to AI adoption.

This growing trend suggests companies no longer view AI as an optional tool. It is increasingly becoming a core layer of workplace productivity.

The rise of AI generated code often raises fears about job losses, but the reality may be more complex. Rather than replacing engineers outright, many firms appear to be redesigning engineering roles.

Routine coding, repetitive debugging and documentation may increasingly be automated. Meanwhile, demand could rise for professionals skilled in architecture, security review, product judgment, AI oversight and system integration.

Developers who can collaborate effectively with AI systems may become more valuable than those relying only on traditional workflows.

For younger professionals entering the field, coding knowledge remains important, but broader skills such as product thinking, logic, communication and quality assurance may matter even more in the coming years.

When the CEO of Google says three quarters of new code is now AI generated, it sends a powerful signal across the global technology industry.

Google runs some of the world’s most advanced infrastructure, from search and cloud computing to Android, YouTube and enterprise platforms. If AI can meaningfully accelerate development at that scale, other companies are likely to follow quickly.

The announcement also raises bigger questions about trust, accountability and the future of human work in software engineering. Businesses may gain speed, but they will also need stronger governance to ensure AI generated systems remain safe and reliable.

Google’s latest disclosure shows that AI inside corporations is moving beyond experiments into core production systems. Code writing, migration, marketing and operational tasks are already being reshaped.

For now, humans remain firmly in the approval loop. But as AI systems improve, the relationship between engineers and machines will continue to evolve.

Pichai’s message is ultimately about more than one statistic. It is about a new era where software may increasingly be designed by people, drafted by machines and refined through collaboration between both.