PARIS – France published its first official guide for public sector employees on artificial intelligence use, establishing a framework for responsible implementation across government operations.
The inter-ministerial document aims to balance innovation with data security and accountability, setting a precedent for comprehensive guidelines in public administration.
The guide outlines five fundamental principles for AI use: agent responsibility, tool selection based on data sensitivity, transparency, sobriety, and continuous training, according to the document.
Public agents remain fully responsible for content generated by AI tools, which must be validated before any dissemination, the guide stated. It explicitly prohibits the automation of critical actions by artificial intelligence.
For content substantially produced by AI, transparency is mandated, requiring a mention of ‘partially generated by artificial intelligence’.
The document categorizes AI tools into three types. Highly secure public administration tools, such as the inter-ministerial AI Assistant, are at the top tier. Commercial paid tools with enhanced guarantees form the second category.
Free commercial tools, including platforms like ChatGPT, are subject to very restricted or prohibited use, particularly when handling sensitive data, the guide reported.
The publication responds to significant demand from public agents seeking clear rules for AI usage, according to David Amiel, an official involved in the initiative.
The guide is part of an ongoing social dialogue process and may evolve following negotiations with trade unions, officials said.
The French government’s move underscores a commitment to integrating AI into public services while ensuring ethical considerations and data protection are paramount. Ministries such as National Education and Agriculture are expected to integrate these guidelines into their operations.
Source: Abondance
