CDT Continues to Oppose Federal Moratorium on State AI Laws
In late May, the U.S. House of Representatives passed their version of the budget reconciliation bill including a provision that would prevent, for ten years, states and localities from enforcing any laws that protect people against harms caused by artificial intelligence (AI) systems. Revised versions have been circulated as the Senate considers the bill, some of which made the moratorium worse. On the whole, CDT believes there is no acceptable version of an “AI moratorium,” as we and dozens of other groups pointed out Wednesday in a letter to Senate leadership.
The most recent version of the provision now conditions states’ access to Broadband Equity Access and Deployment (BEAD) funding on whether they accept the moratorium. As CDT’s Eric Null argued, “Even if the AI moratorium were narrowed, it’s still a bad policy that would force states to choose between accepting more funds to get communities online and protecting their residents from AI harms for 10 years. AI protections and broadband infrastructure buildout should happen hand-in-hand, not be an either-or choice. AI companies in the states that take more BEAD funding will have no accountability for their practices under either state or federal law. Congress should eliminate the provision from the reconciliation bill.”
In a letter opposing the moratorium earlier this month, CDT and over 50 other groups highlighted how automated decision-making systems frequently exhibit systemic biases against marginalized communities, meaning that the provision would have particular impact on marginalized people and could even prevent enforcement of state civil rights laws. CDT’s Ariana Aboulafia and Travis Hall also discussed in an op-ed how the moratorium would harm people with disabilities.
In Wednesday’s letter — and in our initial letter opposing federal preemption of AI laws — we argue that, because Congress has not enacted comprehensive legislation enshrining AI protections, the moratorium would leave millions of Americans more vulnerable to the unpredictable safety risks posed by unregulated or inadequately governed AI. Harms from AI systems are well-documented, and in the absence of Congressional action, states’ efforts are necessary to protect people from those harms, allow Americans to reap the benefits of AI, and build trust in the technology.
| |
— CDT and EFF filed an amicus brief in Garcia v. Character Technologies, Inc., urging a Florida federal district court to grant an appeal to ensure adequate review of users’ First Amendment rights. We urge the court to prioritize users’ interests in accessing chatbot outputs in its First Amendment analysis, and argue that failure to consider First Amendment rights as such would open the door for unprecedented governmental interference in the ways that people can create, seek, and share information.
| |
 | AI-generated blue-toned image of security cameras in the foreground and a city skyline in the background.
|
|
— CDT’s Tom Bowman and intern Bhavana Bheem wrote that an untargeted facial recognition system in use in New Orleans puts innocent people at constant risk of being misidentified and arrested for crimes they did not commit. The program — the product of an under-the-table partnership between the New Orleans Police Department and private nonprofit Project NOLA — enables facial scanning of countless individuals, circumventing the city’s own laws and subjecting residents to known accuracy issues with facial recognition technologies.
— We discussed the implications of a judicial order in New York Times Co. v. OpenAI that would require preservation of hundreds of millions of ChatGPT users’ most sensitive conversations, on the theory that some of them might include the New York Times’ copyrighted material. We argue that this step by the court would risk the privacy rights of millions of ChatGPT users — and, given that broad data retention schemes are too risky even when used to combat crime, they are certainly too dangerous to impose in a copyright infringement lawsuit.
— Our Equity in Civic Technology team wrote about how government agencies have for years used AI tools to target specific instances of fraud, waste, and abuse, and why their tactics are more effective than the Trump Administration’s sledgehammer approach. As we’ve already seen, DOGE’s effort to compile massive amounts of data and feed it into AI systems often does not account for the data’s quirks and context, thereby leading to errors.
| |
— CDT’s Elizabeth Laird was quoted by NPR discussing a recent transfer of state Medicaid data to the U.S. Department of Homeland Security (DHS): “This is exactly why privacy advocates have been sounding the alarm about this administration's unprecedentedly boundless approach to collecting people’s data,” she said. ”Not only will this discourage states from cooperating with the federal government going forward, it will lead people to miss out on life-saving benefits they’re entitled to as they can’t trust that their information won’t be used against them.”
— CDT’s Jake Laperruque and 404 Media discussed airlines’ sale of data to DHS: “While obtaining domestic airline data—like many other transaction and purchase records—generally doesn’t require a warrant, there’s still supposed to go through a legal process that ensures independent oversight and limits data collection to records that will support an investigation. As with many other types of sensitive and revealing data, the government seems intent on using data brokers to buy their way around important guardrails and limits.”
— CDT’s Kate Ruane was quoted by Time Magazine on disinformation surrounding protests in Los Angeles: “The dysfunctional information environment we’re living in is without doubt exacerbating the public’s difficulty in navigating the current state of the protests in LA and the federal government’s actions to deploy military personnel to quell them,” she said.
— CDT’s Ariana Aboulafia and Towards Justice’s Nina DiSalvo coauthored an op-ed at Tech Policy Press about how surveillance wage-setting can affect people with disabilities.
| |
 | Event graphic for CDT's 2025 Annual Benefit, Tech Prom. Abstract gradients of dark blue, and dark blue and gold text.
|
| CDT "In-Person"
— CDT is pleased to announce our annual Tech Prom on Thursday, October 23, 2025, at The LINE DC. Join us for a fun night of networking and conversation — you won’t want to miss it! Sponsorships and individual tickets are available now.
— This week, CDT’s Kate Ruane and Dhanaraj Thakur attended the Internet Governance Forum. Kate spoke on the panel “Protecting Children from Online Sexual Exploitation, including in livestreaming spaces,” and Dhanaraj appeared on panels to discuss generative AI and large language models in content moderation.
| |
 | Image of Isabel Linzer wearing a navy shirt. CDT logo in the background.
|
| Staff Spotlight
How long have you been working in digital rights?
My first experience working on digital rights was at Freedom House in 2018, when I began coordinating sub-Saharan Africa research for the Freedom on the Net report. Since then, protecting human rights online has been a constant through my work, whether on elections or transnational repression or while pursuing my master’s degree.
What is your proudest moment while here at CDT?
I was delighted to represent CDT at the Council of Europe’s Octopus Conference earlier this month. In addition to being reminded of the concrete policy progress taking place around the world, it was rewarding to share the work and immense knowledge developed by the CDT team, from the role of AI in elections to concerning trends in deepfakes legislation.
What is the best book you've read recently?
I cannot recommend "My Friends" by Hisham Matar highly enough. The protagonist is a Libyan student who lives as an exile in London through adulthood. The writing is beautiful, and while we haven’t all lived in exile in fear of Gadaffi’s regime, searching for a place and for people that feel like home is something anyone can understand. To see Matar reflect human experience so specifically and universally at once is a reminder of the power of empathy and, therefore, why we need novels.
What is your fandom?
I’m a lifelong sci-fi and fantasy fan (shout-out to my dad for reading Lord of the Rings to me when I was in kindergarten). Podcasts about the stories I love have been a huge part of me enjoying genre and connecting to the fandom as an adult. In particular, I’m a loyal listener to "House of R" from The Ringer. If you want semi-unhinged, super-smart three-hour podcasts about shows and movies you’re watching, it’s the podcast for you. P.S. We’re currently reading "Babel" for CDT book club, so I feel very lucky to enjoy my fantasy fandom at work, too!
| |
|
|
|
|
