Microsoft is now utilizing artificial intelligence to write a significant portion of its code. Currently, AI generates as much as 30% of Microsoft’s code, according to company heads. This adoption of generative AI coding tools marks one of the technology’s first major business applications and is rapidly changing software production.
Generative AI Now Codes 30% of Microsoft & Google’s Software
Across major tech companies, generative AI is increasingly responsible for creating software code. Google currently has AI generating over 25% of its code, while Microsoft reports AI handles as much as 30% of its codebase. These figures were shared by the leadership of both companies, demonstrating a significant shift in software development practices and an early, practical application of the technology. The rise of AI coding tools—including GitHub Copilot, Cursor, Lovable, and Replit—is also impacting the job market. While these tools help existing software engineers, researchers at MIT CSAIL note AI-generated code isn’t always reliable or secure. This adoption could lead to a reduction in entry-level coding positions as AI handles more routine tasks.
GitHub Copilot, Cursor, and the Rise of “Vibe Coding”
GitHub Copilot, Cursor, Lovable, and Replit are examples of powerful new AI tools enabling individuals with limited coding experience to create digital projects. These tools utilize generative AI to produce, test, and debug code based on user prompts, streamlining the development process. A new coding approach called “vibe coding” is emerging, where practitioners accept and integrate a significant portion of the AI’s code suggestions into their projects. However, reliance on AI coding assistants isn’t without caveats, as the technology can generate inaccurate or insecure code.
Researchers at MIT CSAIL caution that even plausible-looking AI-generated code may not function as intended, highlighting the continued importance of human expertise. Furthermore, early impacts of these tools include a potential reduction in entry-level coding positions within the industry.
AI-Generated Code Risks: Hallucinations and Job Displacement
AI coding tools, despite aiding current software engineers, present risks regarding code reliability. Researchers at MIT CSAIL have found that these systems “hallucinate nonsense,” meaning the generated code may appear correct but doesn’t function as intended or could create security vulnerabilities. This lack of guaranteed functionality emphasizes the continued need for human expertise to review and validate AI-generated suggestions. The increasing adoption of AI in coding is also impacting employment opportunities, particularly for those starting out. Early indications suggest a reduction in entry-level coding jobs as AI tools automate some of the more tedious tasks previously assigned to junior developers. While these tools enhance productivity for existing professionals, they may limit pathways for new workers entering the field.
AI now writes as much as 30% of Microsoft’s code and more than a quarter of Google’s , according to the heads of those companies, while Mark Zuckerberg aspires to have most of Meta’s code written by AI agents in the near future.
