Software Engineers Retool as AI Writes 75% of Google Code
Updated
Updated · The Guardian · Jul 13
Software Engineers Retool as AI Writes 75% of Google Code
1 articles · Updated · The Guardian · Jul 13
Summary
More engineers are shifting from writing software to reviewing AI-generated code, prompting many to sharpen core coding skills, learn AI oversight, or consider leaving the field.
Google says AI now writes 75% of its code, while experts say engineers' value is moving toward defining problems, designing systems, and checking AI output for bugs, security flaws, and weak logic.
Job anxiety has deepened as more than 600,000 US tech workers have been laid off since ChatGPT's 2022 release, computer science graduate unemployment hit 7% in 2024, and tech job postings on Indeed fell 36% from 2020 to 2025.
Some workers are finding paths back in by building AI skills—one laid-off engineer landed an AI-focused role after nearly two years and 400 applications—while others are organizing for protections through new worker groups and union efforts.
The shift is also cooling the profession's pipeline: enrollment in four-year computer and information science programs fell 8.1% in the 2025-2026 school year, underscoring how AI is recasting coding from a prized craft into a supervised workflow.
If AI automates entry-level work, how will the next generation of expert engineers be trained?
With AI's soaring costs and security flaws, is replacing human coders a massive corporate miscalculation?
When AI writes the code, who is liable for the bugs, security breaches, and copyright lawsuits?
The 65% AI Code Revolution: Industry-Wide Transformation, Workforce Upheaval, and the Future of Software Engineering
Overview
The software development industry is experiencing an unprecedented transformation, driven by the rapid ascent of AI-generated code. This shift is fundamentally redefining how software is built, with AI now responsible for producing most new code at leading companies like Google. As a result, engineers are moving from being primary code writers to supervisors of automated systems, leading to significant productivity gains and changes in workforce planning. Companies such as Snap have already seen over 65% of new code generated by AI, prompting reductions in planned headcount and highlighting the profound impact of AI adoption across the industry.