Traditionally, software development was grounded in rigor, discipline and established methodologies such as Waterfall and Agile. These approaches emphasized structured requirements gathering, design validation, code reviews, testing cycles, formal change control, etc. The policies, standards and frameworks were followed by development teams.
In contrast, “vibe coding” represents a paradigm shift where software is created through high-level natural language “vibes” rather than manual line-by-line programming language syntax. Reliance is placed on intuition, rapid experimentation, and AI-generated outputs without sufficient validation. This in turn introduces significant risks: ownership, IP, data privacy, copyright, attribution and authorship which may expose organisations to legal liability.
Since most coding decisions happen at a technical level, it is therefore incumbent upon the Board to ensure that AI policies and governance structures are in place to guide the organisation not only in terms of which AI tools are allowed, but also the developer-level guardrails for compliance, reputational and legal purposes.