To view this email online, paste this link into your browser:
[link removed]





AUGUST U.S. NEWSLETTER 

([link removed])

Pushing Back Against Federal Policies That Would Enable AI Discrimination

In recent months, the federal agencies charged with protecting U.S. consumers and students have made it harder to curb discrimination that can be caused by AI systems. Various policy changes will regress civil rights enforcement, and would chill AI developers from implementing essential safeguards — and CDT is speaking up at this critical moment for AI accountability.

Because proving discrimination often requires intent, legal recourse for the decisions an AI system makes can be hard to come by. It’s nearly impossible to prove that an AI system was deployed to intentionally discriminate, when consumers have little visibility into how the system was built or used. That’s why the “disparate impact” legal theory is instrumental in demonstrating harm created by an AI system: it focuses on outcomes, and whether a neutral practice or policy has a consistent, adverse, and disproportionate impact on members of a protected group. Despite the theory’s importance in a world where AI systems’ decisions shape how millions of people live their lives and access opportunity, the federal government is stepping back from using the concept to enforce civil rights. 

Earlier this month, when the FTC announced that it will no longer pursue disparate impact claims, CDT’s Ridhi Shetty pushed back in a statement to media: “Contrary to the FTC’s assertions, disparate impact liability serves a critical purpose. It requires companies to do basic due diligence to ensure their policies and practices are fair to everyone they interact with. Protecting against disparate impact is key to ensuring that companies comply with laws that protect equal opportunity in the U.S., like the Equal Credit Opportunity Act or the FTC Act, especially as AI is increasingly used in making fundamental decisions subject to these laws.” The FTC indicated that it will focus exclusively on intentional discrimination, which, as Shetty further explained, will “make advancing discrimination claims involving AI incredibly difficult.” 

In late July, the Department of Education — one of the first federal agencies to put out guidance on the intersection of civil rights and AI — rescinded its own guidance on disparate impact. CDT’s Kristin Woelfel, who recently detailed AI’s impact on gaps in student achievement and opportunity ([link removed]), argued that the move “undermines the moral imperative to ensure that all students are given the opportunity to succeed. The body of research showing that well-intended uses of AI in school settings are unintentionally harming students is growing, and parents are demanding accountability from their children’s schools. This is the worst moment to repeal safeguards that could protect children from lasting harms that result from unintended consequences of this technology.” 

CDT has also called on the Consumer Financial Protection Bureau ([link removed]) (CFPB) to reverse course on its proposal to eliminate disparate impact liability from a key regulation, and preserve crucial protections for consumers against credit discrimination — particularly when AI is involved.

AI developers aiming to prevent discrimination — and implement other safeguards — are under threat too, specifically as part of the federal government’s broad campaign to police private expression that does not reflect its preferred political and social ideology. A proposed FTC policy ([link removed]) purports to prevent “suppression of accuracy” in AI systems, but in fact wrongly treats standard technical steps ([link removed]) to ensure AI models are accurate, reliable, fair, and safe as potential “deceptive steering” away from a nonexistent “neutral” baseline. In formal comments to the FTC, we warned that the proposal would chill developers ([link removed]) from implementing essential safeguards, though an AI developer’s design decisions constitute protected editorial judgment under the First Amendment.

Overall, these policy changes will create a less equitable society, disincentivizing entities including companies and schools from putting AI safeguards in place that help avoid discrimination. As Woelfel says, “If those considerations fall away, there is less pressure on developers to rigorously test and improve their systems, all while these emerging AI systems remain poorly understood." 

In Case You Missed It

—CDT’s State Engagement Team reviewed state and federal trends in AI bills ([link removed]) across the 2026 legislative session, and looked to what’s ahead in 2027 for state AI regulation.

— In a letter from civil society organizations ([link removed]) and comments to the FCC ([link removed]), CDT explained that an FCC measure purporting to address robocalls would impose severe and unjustified privacy costs on hundreds of millions of Americans, as phone companies would have to collect a government-issued ID number, physical address, and alternate telephone number from every new and renewing subscriber. We urged the FCC to instead focus its robocall efforts on high-volume callers and foreign-based customers.  

([link removed])

Illustration of three white cameras facing different directions on a gradient blue background. 

— New CDT polling of the UK public ([link removed]) revealed consistent and overwhelming pushback against the country’s expanding surveillance frameworks, which threaten to infringe upon encrypted communications. The results suggest that the British public have deep reservations about state access to their private conversations. “If there is a warrant process in place that notifies the user, or at least provides an opportunity to challenge it, that would satisfy a lot of the pushback we are seeing in polling,” CDT’s Tom Bowman told Computer Weekly ([link removed]). 

— A bill that begins to address several of the core health privacy protections that CDT has advocated for passed a Senate HELP Committee markup ([link removed]). While the bill as amended is not perfect, we look forward to working with lawmakers to further improve the bill.

— Following a hearing earlier this month on surveillance pricing, CDT commended the Senate Judiciary Committee’s Subcommittee on Crime for investigating the practice of using a consumer’s personal digital profile to gauge the opportunity to charge them more. We argue that the technology is increasingly available and powerful thanks to AI ([link removed]), and should be reined in because it’s unfair, discriminatory, and violates privacy.

— Spread the word: CDT is now accepting applications ([link removed]) for our 2027–2028 class of non-resident fellows! The program ([link removed]) is an opportunity for academics from a wide variety of disciplines to engage directly with our policy work and help translate research into actionable insights. 

([link removed])

Digital poster for the 2026 Future of Speech Online Event featuring a speech bubble and white text on a black background.

CDT "In Person"

— Register now ([link removed]): From September 23-24, CDT’s annual Future of Speech Online event ([link removed]) will examine the need for online anonymity, where it should or shouldn’t be compromised, the emerging and growing threats to anonymity online, and what it takes to support anonymous speech into the future.

— Our most recent Tech Talks episode is a special live feature! CDT’s Miranda Bogen interviewed Ami Fields-Meyer and Julia Angwin about their newly published book, On Courage: How to Be a Dissident in an Age of Fear. ([link removed])

— On September 9 at the Ars Electronica Festival for Art, Technology, and Society ([link removed]) in Linz, Austria, CDT’s Isabel Linzer joins a panel, “From Outrage to Influence: How Psychology, Creators and Platforms Shape Political Debate. ([link removed])”

([link removed])

Graphic for CDT's Tech Prom. Blue and white text on a striped black and blue background.

— Don’t miss CDT’s 2026 Tech Prom ([link removed])! Join us at the LINE DC on Thursday, November 12, 2026 for a fun night of networking and conversation — you won’t want to miss it! Sponsorships and individual tickets are available now.

— On September 28, CDT and GW Law will host a half-day symposium ([link removed]) on the real-world challenges and impacts that speakers of some languages face when using multilingual AI technologies. In-person capacity is limited, and advance registration is encouraged.

— On September 22, CDT’s Becca Branum will speak at a Hill briefing ([link removed]) on the emerging First Amendment questions raised by AI systems, including chatbot outputs, transparency mandates, model restrictions, and civil rights enforcement.

CDT In the News

— CDT’s Samir Jain spoke with NPR ([link removed]) about the recently announced settlement between Meta and many states over allegations about how the company’s products affect children. CDT’s Kate Ruane was also quoted by Fortune on the settlement’s impact ([link removed]): "Giving kids and their parents more choices and control is beneficial," she said. "We also see the potential for significant risks to everyone’s privacy and free expression rights online, especially in the ways this settlement will subject all users to invasive age assurance and limit all kids’ access to content and services regardless of their individual needs."

— CDT’s Tom Bowman was quoted by Good Morning America ([link removed]) about changes announced this month by Flock, a provider of automated license plate reader technology, in response to public pressure: "While this measure contains some meaningful improvements, such as the shortened retention policy, ultimately it is the same 'Just Trust Us' policy with a fresh coat of paint," he said. "Flock is still asking thousands of police departments to police themselves, despite evidence from across the U.S. that this lack of guardrails is unsafe.” And, at Tech Policy Press, CDT’s Jake Laperruque wrote an op-ed ([link removed]) arguing that unless Flock is reined in, mass video surveillance will grow even more pervasive.

— CDT’s Travis Hall joined NPR’s Weekend Edition ([link removed]) to discuss bans on sales of people's geolocation data, in light of a Maryland law that took effect July 1 stopping data brokers from providing sensitive data to entities that work with immigration enforcement.

— CDT’s Miranda Bogen spoke with Axios ([link removed]) about the data that AI chatbots collect and what happens to it: "The more a system knows about you, the easier it will be to make escalating requests for private details in a way that feels natural," Bogen said. "Without robust privacy protections, the incentive to monetize that knowledge will be hard to resist."

— CDT’s Greg Nojeim spoke with Mashable ([link removed]) about a recent court ruling finding that the Trump Administration’s designation of Anthropic as a supply chain risk constituted unlawful retaliation: "Yesterday’s ruling is a vindication of the First Amendment and a clear reminder that the government isn’t allowed to use national security designations to punish companies or individuals for public criticisms or business disagreements," he said. "Procurement policy can’t be an end run around the Constitution. The Pentagon can pick and choose which products it wants to buy. It can’t use the power of its size and position to stamp out disagreement."

([link removed])

Nick Doty smiling wearing a dark blue shirt with a green background.

Staff Spotlight

Nick Doty ([link removed]), Senior Technologist

How long have you been working in digital rights?

When I left software engineering to go back to grad school in 2008, I initially thought my interest was the abstract challenges of organizing and analyzing data. But I quickly discovered that tech policy issues like online privacy were both meaningful and fascinating, because of the messy, ethical, human questions involved. I'm so grateful to Professor Deirdre Mulligan (former CDTer) for introducing me to those questions, and I've been working on human rights on the internet ever since.

What do you want people to know about your area of focus in tech policy?

Technical standard-setting can be arcane and jargon-filled, but it's an essential part of what makes the internet so distinctive. People from different backgrounds, companies, universities, and occasionally civil society organizations, debate protocols and how online software should interoperate. I work most often (and studied for my dissertation) in the World Wide Web Consortium and the Internet Engineering Task Force. The decisions those groups of engineers make have profound impacts, good and bad, for our freedom and our privacy in communicating online.

What is the most recent cultural activity you’ve been to?

As a parent, the Saturday farmer's market here in Durham is the social and cultural focus of the week: fresh food, live music, arts and crafts, friends and family, and a huge playground.

What is the best book you've read recently?

I was so impressed with To the Lighthouse, a classic I only picked up recently. After too much time scrolling and quickly reading things on my phone, it took me a while to read the long stream of consciousness, but I was moved by the family's interactions and the artist's outsider perspective.

#CONNECT WITH CDT

SUPPORT OUR WORK ([link removed])

([link removed])

([link removed])

([link removed])

([link removed])





1401 K St NW Suite 200 | Washington, DC xxxxxx United States

This email was sent to [email protected].
To ensure that you continue receiving our emails,
please add us to your address book or safe list.

manage your preferences ([link removed])
opt out ([link removed]) using TrueRemove(r).

Got this as a forward? Sign up ([link removed]) to receive our future emails.
email powered by Emma(R)
[link removed]