AI Policy for Safe Company Use

The Internal AI Policy for Safe Company Use is a practical, adaptable policy framework published in 2026 by Masonsoft Technology Ltd. It is designed for organisations that are already using AI tools or about to adopt them. It provides a complete governance structure that can be scaled to fit businesses of any size. The document is organised into seven substantive sections. The opening section establishes four guiding principles covering transparency, data protection, human accountability, and fitness for purpose. The second section sets out hard prohibitions on data confidentiality, the treatment of AI output, high-stakes decision-making, and the misrepresentation of AI-generated work. The third section covers positive obligations, including tool approval processes, input sanitisation, output verification, and process documentation. A dedicated section for managers addresses expectation-setting, modelling good practice, and recognising overreliance. A risk register covers five specific risk categories with practical mitigation guidance for each. A further section provides strategic direction for senior leaders and boards, covering governance ownership, regulatory developments, and the ethical dimensions of AI use. The document includes a quick-reference card for day-to-day use and adaptation notes that make it straightforward to customise for different organisational contexts and sizes.