“Vibe coding” named word of the year – highlighting AI’s growing influence

Collins Dictionary, one of the most influential English language dictionaries, selected “Vibe coding” as its 2025 Word of the Year. The term describes a new method of software development that uses artificial intelligence to convert natural language into computer code, allowing people with little or no programming experience to write code.

Collins’ editors monitor language trends using the 24-billion-word Collins Corpus, which gathers content from various sources, including social media and news outlets. Each year, they choose emerging and influential words that best reflect changes in modern language.

The decision to highlight “vibe coding” came after a sharp rise in its usage since its introduction in February. The phrase was coined by Andrej Karpathy, former Director of AI at Tesla and a founding engineer at OpenAI, who used it to describe how AI enables users to create applications while effectively “forgetting that the code even exists.”

Vibe coding wasn’t the only term highlighted by Collins. The list also included other terms such as “biohacking,” which refers to modifying one’s biological functions to improve health and longevity, and “clanker,” a negative term for robots, computers, or AI systems. The latter became popular online as a way for users to express irritation or mistrust toward AI platforms and chatbots.

Since AI chatbots were introduced in late 2023, they have caused a major shift in many professions, with programming being one of the most affected. “Vibe coding” now enables junior programmers, as well as individuals with no coding experience, to produce mostly operational code.

Surely, code written by AI is still far from perfect. It has numerous errors and security issues, which are usually addressed by engineers overseeing it. However, this has caused anxiety because writing flawed code used to be an essential learning experience for junior programmers. Meanwhile, experienced programmers fear being assigned only to code review tasks, raising production targets and leaving them with their least favorite duties.