Center for Democracy & Technology

30 years strong thanks to the support of our community! For the past three decades, the Center for Democracy & Technology has been at the forefront of defending civil liberties and advancing digital rights in an ever-evolving technological landscape. Since our founding on December 6, 1994, we’ve worked tirelessly to protect freedom of expression, privacy, and innovation online. As we celebrate this milestone, we’re so grateful for CDT’s founders, alumni, and the broad community that has shaped CDT’s journey and impact. Together, we’ve tackled some of the most critical issues in #TechPolicy, from free speech online to digital #privacy, and have been there for the defining moments in internet history: ✨ 1996: CDT led efforts against the Communications Decency Act and successfully argued before the #SupremeCourt in Reno v. ACLU. ✨ 2001: In the wake of 9/11, CDT stood strong, advocating for privacy and #CivilLiberties against the sweeping #surveillance provisions of the USA PATRIOT Act. ✨ 2004: CDT embraced the growing role of the internet in politics, launching Youth04, a grassroots campaign encouraging young voters to engage in the 2004 election.  ✨ 2009: As people shared ever greater portions of their lives online, CDT launched Take Back Your Privacy, a campaign to mobilize support for comprehensive federal privacy legislation and establish clear rights for consumers and clear rules for companies that collect personal information. ✨ 2013: In response to Edward Snowden’s revelations, CDT continued to push for legislative reforms to strengthen digital privacy and security. ✨ 2018: CDT helped establish the Santa Clara Principles to set guidelines for transparency and accountability in #ContentModeration on social media platforms—standards that continue to influence global tech policy. ✨ 2022: With the rise of #AI, CDT was at the forefront of policy efforts in the U.S. and EU to advance responsible, rights-respecting innovation – and launched the AI Governance Lab as a leading source of public interest technical expertise. CDT’s leadership in these and many other moments of tech history wouldn’t have been possible without the contributions of our incredible community. Thank you for your dedication and partnership over the years! Keep an eye out for more details on how you can stay involved as we celebrate our 30th Anniversary throughout 2025! cc: Jerry Berman, Nuala O'Connor, Lisa A. Hayes, Christopher Calabrese, Alexandra Reeve Givens, Ari Schwartz, Alan Davidson, Deirdre Mulligan, James Dempsey, Jonah Seiger.

🌸 CDT’s Spring Fling is back! 🌸 Join us on Tuesday, April 22, 2025, 8–10:30 PM at Dirty Habit in D.C. for a night of connection and celebration alongside IAPP’s Global Privacy Summit. Over 500 privacy & AI leaders will be there—save the date! Sponsorships available with perks like brand recognition & tickets. Reach out to learn more!

🎙️ New CDT Tech Talk: What happens when AI decides to “resign”? CDT’s Jamal Magby is joined by Umang Bhatt, CDT Non-Resident Fellow, to explore how stopping recommendations or restricting outputs could mitigate reputational & legal risks. Listen Now! https://lnkd.in/e2kzSzhA

“Meta’s civil rights advisory group sent a letter Tuesday to Meta CEO Mark Zuckerberg to voice their ‘grave concern’ over Meta’s recent content moderation changes.” https://lnkd.in/e8EZ9Kvv Read the full letter: https://lnkd.in/eB9rHQ9A

Yesterday, President Biden signed a significant executive order in support of cybersecurity. The order outlines a number of smart, constructive steps to protect government systems and the sensitive data they contain. https://lnkd.in/eg8pNSmX End-to-end encrypted communications, phishing-resistant credentials, software supply-chain information and secure Internet routing are all important improvements to protecting government systems and data. The EO also includes a commitment to minimize the risks of fraud and waste while maintaining access to public benefits through the use of digital identity technologies. The initial steps in this order lay out a path towards digital identity while supporting principles of privacy and interoperability. Those principles are important to ensure that digital identity documents secure access to public benefits and services but don’t contribute to broader surveillance and exclusion or make it unnecessarily hard to access life-saving benefits.

Earlier this week, CDT submitted comments to NTIA in response to its request for comment on ethical guidelines for research using pervasive data. CDT’s comments emphasize the importance of the ethical use of pervasive data through independent research to inform the governance of technology. Read the full comments: https://lnkd.in/eudaTdvG

New explainer from CDT’s Amy Winecoff + Miranda Bogen dives into the fundamentals of hypothesis testing, how auditors can apply it to AI systems, and where it might fall short. Using simulations, we demonstrate how hypothesis tests could succeed (and fail) in detecting bias in a hypothetical hiring algorithm. Hypothesis testing is a key tool for AI auditors, offering a rigorous way to evaluate and interpret data. It quantifies uncertainty, aiding informed decisions and actionable insights. But like any tool, it has its limits. In AI audits, hypothesis testing can uncover disparities, like gender bias in hiring algorithms. But results can be affected by randomness, like the data subsets chosen. Understanding these limitations is crucial for valid, actionable findings. By understanding both the strengths and limitations of hypothesis testing, AI auditors can make smarter, more nuanced recommendations. Explore how it works and why it matters in Amy Winecoff + Miranda Bogen's explainer. https://lnkd.in/ek-ZSwFU

Our “Education Leader’s Guide” explores how immigration enforcement may impact schools and provides actionable recommendations to safeguard students and their families. https://lnkd.in/eBH6Zzcy cc: Kristin W. With immigration enforcement likely to intensify this year, school administrators must comply with privacy and civil rights laws when collecting data and using technology. While ICE traditionally avoids enforcement actions on K-12 campuses, this is a norm, not a prohibition. Schools must be prepared to address potential enforcement on campus. Schools can and should uphold long-standing legal obligations that protect student privacy and rights, regardless of any shifts in immigration enforcement policies. https://lnkd.in/eBH6Zzcy

As AI adoption grows, so does the need for accountability. Traditional methods like audits & benchmarks often focus on narrow technical outcomes, but what about broader social impacts? CDT’s latest blog by Evani Radiya-Dixit, explores how holistic, sociotechnical AI assessments can fill the gap. AI systems are embedded in society, impacting human relationships & institutions. Assessments must look beyond technical components to consider: 1️⃣ Technical layers: data, inputs, & outputs 2️⃣ Human interaction: how users engage with AI 3️⃣ Systemic impacts: effects on labor markets & rights To truly understand AI’s risks & impacts, we need sociotechnical frameworks that connect the technical with the societal. Holistic assessments can guide responsible AI deployment & safeguard safety and rights. 📖 Read more: https://lnkd.in/eD7mDzd8

NEW REPORT: CDT AI Governance Lab’s Assessing AI report looks at the rise of complex automated systems which demand a robust ecosystem for managing risks and ensuring accountability. Discussions often get muddled by overlapping terms like auditing, impact assessments, red-teaming, and assurance. Tools like audits, impact assessments, & red-teaming help identify risks, but they differ in scope and methods. CDT’s “Assessing AI” report aims to cut through the fog, focusing on shared goals: ✅ Identifying risks ✅ Evaluating adequacy ✅ Communicating impacts ✅ Incentivizing change Aligning on these objectives will help refine and enhance accountability methods.From mapping policy proposals to examining scope & independence, we identify steps to build a mature, inclusive risk management ecosystem. Clear goals = effective implementation. Read the full report from CDT AI Governance Lab Director Miranda Bogen: https://lnkd.in/eV29APwd