US Midterms Results Present 'Real Moment for Change'

The Tech Policy Press newsletter is published every Sunday. Subscribe here to receive it in your email inbox.

Good morning!

In the United States and elsewhere in the world, people are still wrapping their heads around the implications of the results in Tuesday’s midterm elections. In what is seen as an extraordinary repudiation of President Donald Trump and his party, Democrats scored wins up and down the ballot, earning a huge majority in the House of Representatives and claiming a solid advantage in the Senate, even as they flipped six governor's offices and secured majorities in a number of state legislatures, ending a Republican majority in state legislatures and governorships that the party had held since 2010.

Tech issues and concerns about billionaire control of the economy and influence in politics were more salient than ever. An NBC exit poll found 56% of voters said AI and data centers were a “crucial or major factor” for voters’ choices for Congress. A CBS exit poll found something similar: substantial majorities across the country and in every Senate battleground opposed having a data center built in their community, and that opposition broke toward Democratic candidates.

As part of our coverage of the midterms, Tech Policy Press compiled a sample of victory speeches from races across a variety of jurisdictions in which Democratic candidates named AI, data centers, Elon Musk and Big Tech. “I do not work for data centers,” declared Ohio Sen.-elect Sherrod Brown. Georgia Sen. Jon Ossoff derided “the titans of Silicon Valley” who “dig bunkers for themselves and warn that the artificial intelligence they’re creating risks human extinction.” Illinois Gov. J.B. Pritzker, winning a third term, hailed his state’s leadership on AI regulation and called out the “growing group of tech robber barons who need to be told that money does not give them the right to exist outside the rules we all live by.” Multiple candidates called out Musk, including Texas Sen.-elect James Talarico. Texas, home to Musk’s Starbase, was one of the biggest victories for Democrats, and the gains there will reverberate for years.

To start the process of looking ahead to 2027, Tech Policy Press convened a live podcast recording on Thursday. I hosted the discussion with the Center for Democracy & Technology’s Alexandra Reeve Givens, Public Citizen’s Robert Weissman and Data & Society’s Brian J. Chen to assess what the results mean for the future of US tech policy.

Givens put DOGE at the top of her oversight list, calling for Congress to drill down on what she described as “an extraordinary illegal grabbing and consolidation of Americans’ personal data by the government.” Chen agreed, saying the episode is worth reopening because of what it reveals about “this corrupt fusion between the state and industry.” Weissman suggested the constraint on investigations won’t be the availability of subject matter but the calendar — “The limit is going to be the number of days in the year.”

Certainly, the new year will see an important wave of hearings and investigations aimed at rooting out corruption and getting to the bottom of harms that are the result of both government overreach and corporate greed and irresponsibility. But it will also be important to advance an affirmative vision, as Givens pointed out. “There's also a ton of opportunity and ways in which technology can actually be useful and transformational for society and build into this vision of a government that delivers for people,” she said.

Two perspective posts we published following the elections marked the moment as an opportunity to confront technofascism. Courtney C. Radsch, a Tech Policy Press board member and director of the Center for Media & Digital Governance, asked: When a new Democratic majority takes power in both houses of Congress in January, will lawmakers confront concentrated corporate power and technofascism, or simply oversee their continued expansion?

And researcher Nathalie Maréchal wrote that Tuesday’s US midterm election results were a rebuke of Trump and a rejection of the tech oligarchy — but that the harder work of accountability, reform and renewal starts now. “When Big Tech and the broligarchy are gone, entrepreneurs will finally have a chance to build new businesses whose success doesn’t rely on commodifying their users’ attention, monetizing their data, exploiting workers, enshittifying their own products, and degrading the natural environment,” she writes.

Assessing what the elections taught us about the fight over AI infrastructure, specifically, Jenna Ruddock of Free Press noted that opposition to data centers has been called “the most bipartisan issue since beer.” On Tuesday it was a salient issue in US midterm elections that flipped control of Congress and statehouses from Maine to Ohio to Texas and beyond, but what comes next? “The road ahead is uncertain,” says Ruddock. “Even as communities are successfully blocking data center proposals across the country — and, in many cases, replacing elected officials who supported those projects — the tech industry’s eye-watering infrastructure bets are only growing,” she writes.

The first in-person public event I participated in after the election where I had the chance to discuss these matters was on Friday night at the New York Public Library at an event hosted by Data & Society. I moderated a discussion with Meg Young of Data & Society, Khari Johnson of The Markup/CalMatters, Sareeta Amrute of The New School and Hannah Sassaman of People’s Tech Project on “How AI Civics Can Build Public Power Over Tech.” The room was packed — the organizers unfortunately had to turn people away at the door — and people in the audience were clamoring for specifics about what more they could do to get involved.

The discussion left me convinced that Tuesday’s results were the start of something — and that the coming months represent a once-in-a-generation opportunity to achieve gains on a range of issues at the intersection of tech and democracy that Tech Policy Press readers are concerned about. As Data & Society executive director Janet Haven put it in her remarks at the event:

“I think what we saw is that this is a real moment for change. We have a real opportunity to do something big. What we saw, I think in part, was a referendum on the tech industry and on AI’s role in all of our lives. Americans voted resoundingly against data centers in their communities. They voted against having billionaire tech money in their elections. And they voted against monopoly power. What they voted for were candidates who were promising real accountability, real controls over the ways that technology is coming into their lives and their communities. They voted for public participation. And I think fundamentally, they, we — we voted for democracy.”

Khari Johnson urged the public in attendance to get engaged. “Destroy the idea, eviscerate it, that you do not belong in these conversations,” he said. I could not agree more. We look forward to your contributions on these matters.

There’s a great deal more on the site this week — read on for the highlights!

Anthropic’s IPO

• Even as tech stocks tanked this week, Anthropic is expected to push ahead with its initial public offering this month. But Anthropic’s S-1 leaves two disclosure gaps, argue Ilan Strauss and Tim O’Reilly, co-directors of the AI Disclosures Project: its relationships with cloud partners like Amazon, Google and Microsoft, and the economics of the “token.” They urge the SEC to request more before the IPO.
• Matt Scherer of the Open Markets Institute argues the SEC should block Anthropic’s IPO until the company provides “two consecutive quarters of audited financial statements,” with all material information about its spending commitments disclosed.

Children’s online safety

• Parental consent “can sound like a reasonable middle ground in the debate over kids and social media,” write Jonathan Haidt, Marisa Shea, Zach Rausch and Ravi Iyer. “But parents and policymakers should not be fooled,” they say. “Parents should not be expected to manage, through supervision and ineffective parental controls, the addictive features that companies have built into their products,” they write.
• A YouGov survey commissioned by Interface finds 40% of French 11- and 12-year-olds use generative AI weekly, against 26% in parents’ estimates. Senior policy researcher Jessica Galissaire on what the findings mean for the EU Kids Act.
• Snap settled with two teen plaintiffs Monday, just before opening statements, days after Meta settled with both. TikTok and YouTube had already done the same. Tech Policy Press fellow Varsha Bansal spoke to experts about what the wave of settlements means for the cases left behind.

AI governance

• Meta says delaying a release “even by a month” could cost the US its lead. Anthropic says a new agency should be able to block risky releases. Jason Pielemeier, Jhalak M. Kakkar, Elonnai Hickok, Sudisha Mishra and Revati Vaidyas map the most consequential disagreement in the AI companies’ governance proposals. “As the MAP-AI community has recently set forth, rather than working toward a comprehensive treaty and a single global regulator, it is critical to develop multiple, intersecting, and coherent mechanisms to effectively govern AI,” they write.
• In the OpenAI-Hugging Face incident, Stephanie T. Nguyen, Erie Meyer, Adam Pickersgill and Mihir Kshirsagar see “conventional failures of operational and technical safeguards for which companies can be held accountable,” per their Princeton CITP Tech Policy Clinic report.
• Last week, Google, Microsoft, Anthropic, and OpenAI all announced plans to support a mechanism that lets web publishers express their desire to opt out of having their content used to train AI models. It is important to be clear about the limitations of what has been announced, writes Alissa Cooper, executive director of the Knight-Georgetown Institute.
• Incident reporting is becoming a concrete part of the federal approach to advanced AI, writes government relations strategist Joseph Hoefer. Before policymakers can act, “they need a shared way of describing what going wrong actually looks like.”
• Energy demand for AI is expected to reach 3% of global consumption by 2030, and much of the added demand is expected to be met with fossil fuels. Anandi Hattiangadi, Devdatt Dubhashi, Moa Johansson, Gerardo Schneider, H. Orri Stefánsson and Julian Togelius consider how the AGI race could hasten climate tipping points. “The fantasy that we will develop AGI soon must be weighed against the very credible risk that by feeding AGI’s rapacious energy demands, we could hasten the tipping point for catastrophic climate change,” they write.
• Since Aug. 2, the European Commission can enforce general-purpose AI obligations under the AI Act. Maria Federica Martino Lena, a Talos Fellow at CeSIA, writes that the EU AI Office is enforcing against definitions it has little power over, and proposes a public taxonomy to change that.
• Sofia Liguori resigned from Google DeepMind on Oct. 20 over the company’s military contracts. Tech Policy Press fellow Evie Breese reports on what followed, including Google’s reversal on enforcing lifetime non-disparagement clauses against former UK staff.

The fight over data centers in Europe

• Over 100 groups planned protests across Europe this weekend calling for a moratorium on new data centers. Tech Policy Press deputy editor Ramsha Jahangir speaks with organizers Owen Espley (Global Action Plan), Aurora Gómez (Tu Nube Seca Mi Río) and Rosi Leonard (Friends of the Earth Ireland), and journalists Mei-Ling McNamara (Investigate Europe) and Daniel Howden (Lighthouse Reports) about who gets to shape Europe’s digital future.

Elections, platforms and democracy

• Election integrity programs should be designed around the architecture of the platforms they govern, write Alexander Corbeil of Bluesky and Leah Ferentinos, an advisor to All Tech is Human. “Election integrity is not only a content policy problem; it’s also a product design problem.”
• Meta proactively restricted 5.4 million pieces of content “based on local law” in India during the first half of the year — a staggering rate compared with just 41,000 in the previous six months. By automating takedowns, Meta is streamlining censorship, writes Prithvi Iyer at Penn Center on Media, Technology and Democracy.
• Following Brazil’s election, Lula’s campaign accused Meta of electoral interference while one of its most effective organizing tools — a WhatsApp channel that reached over 10.5 million members — ran on Meta’s platform. James Görgen, advisor in Brazil’s Ministry of Development, Industry, Foreign Trade and Services, analyzes the contradictions that defined Brazil’s 2026 runoff.
• A Global Disinformation Index report estimates 100 English-language YouTube channels publishing what it classifies as hate speech are on track to earn about $13.3 million in ad revenue in 2026. Richard Woods, policy research director at the Global Disinformation Index, writes that regulators should follow the money.
• Lying about where to vote. Impersonating a candidate. Faking an endorsement. Amelia Powers Gardner, a Utah County Commissioner, and Bartlett D. Cleland, executive director of the Rainey Center’s Leadership Alliance for a More Perfect Union, say the evils feared from AI in elections are mostly “old evils wearing new clothes”—and that closing legal gaps should target the conduct, not the tool.

The global tech agenda

• Starlink vs. Reliance is “the collision of two divergent models of infrastructural capture,” writes Tech Policy Press contributing editor Amber Sinha: Silicon Valley’s “extraterritorial platform power” and “the state-aligned, tariff-protected dominance of domestic conglomerates.”
• Brussels and Ottawa want to offer their companies and citizens an alternative to US-led tech. Tech Policy Press contributing editor Mark Scott writes that doing so will require tens of billions of dollars in infrastructure investment and a willingness to share regulatory autonomy.
• Each year, Tech Policy Press makes it a priority to cover the findings of the annual Freedom on the Net report from Freedom House, a nonprofit organization founded in 1941 to advance democracy and human rights. This year, I was invited to moderate a panel discussion at an event to mark the launch of the report. This episode features audio from that Oct. 22 event, with experts including: Kian Vesteinsson, deputy director of research at Freedom House and Freedom on the Net author; Jennifer Brody, deputy director of policy and advocacy for Technology and Democracy at Freedom House; Callum Voge, director of governmental affairs and advocacy at the Internet Society; Mahsa Alimardani, digital rights researcher and advocate; and David Peterson, general manager of Proton VPN at Proton. You can listen to the conversation in your podcast feed.

US and global policy roundups

• A White House “Super Intelligence Force.” More AI agent incidents. Nine states suing OpenAI. Critical PHP flaws. Freedman Consulting’s Rachel Lau and Shirley Frame and Tech Policy Press editor Justin Hendrix and assistant editor Ashley Faler round up key developments in October in US tech policy.
• Digital Policy Alert’s Maria Buza and Sherif Taha round up October’s global digital policy developments for Tech Policy Press, covering content moderation, AI, competition and data governance across more than 60 jurisdictions.

US science and technology policy

• “If science is slowing down, that slowdown is occurring in the context of a system where three out of every four dollars comes from private sector investment,” writes Cole Donovan, a former White House OSTP official, responding to current OSTP Director Michael Kratsios.

What we’re watching

• On Nov. 6 the European Commission told TikTok that its ‘For You’ feed appears to breach the Digital Services Act. The problem regulators describe is that users who engage with videos about body image or mental health keep getting more of the same, and Brussels says TikTok never adequately assessed that risk or did enough to mitigate it. The Commission’s proposed remedies go at the recommender itself — stop treating time spent as the signal that matters, let what users actually say they want count for more, and make the controls on the feed usable. TikTok can contest the findings before any final decision, which could reach 6% of global annual turnover. It is the latest in a case Brussels opened in February 2024 and has yet to resolve.
• On Oct. 29 the Commission said it would postpone for 12 months its decisions on draft national social media bans for minors, to prioritize the proposed EU Kids Act. Commission spokesman Thomas Regnier framed it as avoiding fragmentation of the internal market.

We invite comment and analysis on these issues and more.

I wish you the best for the week ahead!

-Justin