A Leading Source of Public Interest Expertise
Launched in 2023, CDT’s AI Governance Lab is a leading public interest voice on how to build, use, and govern AI systems to protect people’s rights and interests. From big questions like personalized AI to emerging issues like AI agents, we provide practical advice on how developers can protect users’ rights while supporting innovation.
In partnership with the Ada Lovelace Institute, CDT published a report on AI assurance, a new field focused on ensuring AI works the way it’s intended. And we’ve advocated for more visibility into AI safety training so researchers and experts can understand how new models actually try to keep people safe. By digging deep on model specifications, training datasets, and evaluation processes, CDT is pushing forward the conversation between companies, regulators, technologists, and civil society to help find solutions that work for all of us.
Helping Government Work Better With AI
CDT is also committed to helping public agencies use AI to deliver better services for people. Our Equity in Civic Tech team has produced a wide range of resources to help governments consider AI use cases with an eye towards protecting rights and privacy. We’ve also provided roadmaps on state legislation and executive actions that govern public sector AI use. Further, we investigated AI governance at the local level, advocating for policies that uphold transparency, accountability, and equity in local and citywide agencies.
Administrators, teachers, students, and parents are on the front lines of AI integration in education. CDT is actively supporting educators by analyzing AI guidance across state education agencies and by offering expert guidance to policymakers. Our original research has helped shine a spotlight on troubling issues like AI plagiarism-detection tools that have disproportionately impacted marginalized students. And we’ve developed model policy and infographics to address incidents of non-consensual intimate imagery (NCII) in K-12 schools that will help school leaders tackle a pressing issue in ways that do more than just punish perpetrators after the harm has been done.
Supporting Lawmakers
Along with our work with local government, CDT is supporting lawmakers at the federal and state level. CDT is closely involved in deliberations over Colorado’s SB 24-205, a landmark law governing the use of AI in consequential decisions that impact people’s lives. To ensure that our “laboratories of democracy” can continue to tackle the tough issues raised by evolving technology in the face of Congressional inaction, we also successfully spoke out against the proposed ten-year federal moratorium on state AI legislation. Following revisions to the moratorium, CDT partnered with more than 140 civil society organizations to send a letter to Senate leadership, outlining how the moratorium would undermine civil rights and civil liberties and cause harm to ordinary people.
Most recently, following the release of the White House’s AI Action Plan and “Woke AI” Executive Order, CDT released new analysis on the practical challenges of defining and measuring ideologically “neutral” AI. We also spoke out against the hypocrisy of the Order’s call for neutrality in large language models (LLMs), given that it also seeks to control the information environment and police LLM outputs.