Link Centre - Search Engine and Internet Directory

Helping to share the web since 1996

From 5% to 29%: The Fast Rise of AI in Programming

Generative AI is rapidly transforming how software is built. According to a new study published in Science, the adoption of AI tools for coding has accelerated dramatically - though not evenly across the world. In the United States, the proportion of newly written code involving AI assistance climbed from about 5% in 2022 to nearly 29% by early 2025. China, by comparison, stood at roughly 12% during the same period.

Computer screen displaying code with a context menu.

This shift matters because software development underpins the modern economy. In the U.S. alone, companies spend hundreds of billions of dollars each year on programming-related wages, while countless lines of code quietly power everything from finance to transportation. The key question is how AI is reshaping this critical sector.

Researchers from the Complexity Science Hub (CSH) analyzed more than 30 million Python code contributions made by around 160,000 developers on GitHub. By late 2024, they found that roughly one in three new software functions written in the U.S. already relied on AI tools such as ChatGPT or GitHub Copilot.

Using a custom-trained detection model, the team identified clear regional differences. While the U.S. led adoption, several European countries followed closely - France and Germany exceeded 20% - with India rapidly closing the gap at around 20%. China and Russia trailed behind, partly due to restrictions on access to leading AI models, though recent domestic breakthroughs suggest this may change.

The study also revealed a divide based on experience. Less seasoned programmers use AI more frequently, yet measurable productivity gains - about 3.6% overall - were seen almost entirely among experienced developers. In fact, senior programmers also used AI to explore unfamiliar libraries and combine tools in novel ways, suggesting the technology enhances learning and creativity rather than just speeding up routine tasks.

Economically, the implications are significant. With U.S. spending on programming work estimated between $637 billion and $1 trillion annually, even modest productivity gains translate into tens of billions of dollars in added value each year - and likely more as adoption continues to grow.

Looking ahead, AI is becoming a foundational element of software development. The challenge for businesses, educators, and policymakers is ensuring that its benefits are broadly shared, rather than reinforcing existing skill and opportunity gaps in the digital workforce.

Newer Articles

Older Articles

← Back to News Headlines