Generative AI-based systems facilitate content creation and enable new forms of communication and expression. Developing fast and at scale, Generative AI can also adversely affect a shared and pluralistic information space, which is essential for democracy.
Adopted by the Council of Europe Steering Committee on Media and Information Society (CDMSI), operating under the authority of the Committee of Ministers, the Guidance Note aims to lay the grounds for a shared understanding of the technology by providing an illustration of the crucial steps of the Generative AI lifecycle. The document also identifies structural implications for freedom of expression, both at an individual and societal level, and delivers a concrete set of actionable measures for policymakers, through an agile governance cycle built on four action areas: observe the impact, assess the risks, enable freedom of expression, and empower users and society at large.
INTRODUCTION - DEFINITION AND SCOPE
1. GENERATIVE AI TECH STACK: FOUNDATION, TOOL AND PRODUCT LAYER
2. FREEDOM OF EXPRESSION AND GENERATIVE AI TECHNOLOGY AND USE
3. GENERATIVE AI STRUCTURAL IMPLICATIONS FOR FREEDOM OF EXPRESSION
3.1. Enhancing expression and content access
3.2. Diversity and standardisation of expression
3.3. Integrity of human expression and its attribution
3.4. Agency and opinion formation
3.5. Media and information pluralism
3.6. Market dynamics
4. GUIDELINES
4.1. Observe
4.2. Assess
4.3. Enable
4.4. Empower
