Business concept of choice: many doors in a field


How AI Gave Me Access—and Why It Must Give Others the Same

A photo of NY Governor Kathy Hochul at her desk signing papers from February 2024.

State AI Rollouts Are Outrunning Their Own Governance

A collage that merges circuit board patterns with textile motifs in a grid-like background of alternating black, grey, and white. Two hand-drawn arms are on each side of the image, positioned as if gently pulling on thin, white strings that cross the image diagonally. The hands appear soft and somewhat translucent, contrasting with the rigid lines of the circuit board patterns behind them. The strings are woven through both the hands and the background, symbolizing the connection between traditional weaving and modern technology. The overall color palette features muted earth tones, including browns, beiges, and grays, creating a sense of both history and continuity between the natural and technological worlds.

Hanna Barakat & Archival Images of AI + AIxDESIGN / 

Project Kaleidoscope: Community Cyber Defense in the Age of AI

CMA Competition and Markets Authority logo on the screen and finger pointing at it. CMA is a government department also known as British competition watchdog.

UK Regulator Staunches Google’s AI Content Grab

Spring garden tours take place on the South Lawn of the White House, Saturday, April 18, 2026. (Official White House photo by Andrea Hanks)

The Open Internet Was Always American. Now It's a Weapon.

Police stand behind their shields as they prepare to clash with protesters outside the Delaney Hall detention center during a protest against the transfer of detainees and federal immigration policies on Saturday, May 30, 2026, in Newark, N.J. (AP Photo/Andres Kudacki)

As Detainees Strike Across the US, Tech Helps Trump Hunt ICE Watchers

Series: What Europe’s AI Market Actually Looks Like

A Fisk University sign is seen July 18, 2025, in Nashville, Tenn. (AP Photo/George Walker IV, File)

The AI Economy Is Coming to HBCUs. Are They Ready for the Tradeoffs?

Henna Virkkunen, Executive Vice-President for Tech Sovereignty, Security and Democracy of the European Commission, at a press conference in Brussels on June 3, 2026. Source: LinkedIn

Henna Virkkunen, Executive Vice-President for Tech Sovereignty, Security and Democracy of the European Commission, at a press conference in Brussels on June 3, 2026. 

EU Unveils Sweeping Tech Sovereignty Push, Balancing Autonomy with Openness

European Commission President Ursula von der Leyen, right, and European Commissioner for Tech Sovereignty, Security and Democracy Henna Virkkunen speak during a media conference at EU headquarters in Brussels, Wednesday, Apr. 15, 2026. (AP Photo/Omar Havana)

How the EU’s Tech Sovereignty Package Finally Puts Open Source to the Test

Cooling pipes hug data servers, extracting water from a shared reservoir while people collect water from the same source, set against a background of eroded soil textures.

The Environmental Impact of Data Centers in Vulnerable Ecosystems by Gloria Mendoza / 

Environmental Lawsuits Present Roadblock for the AI Data Center Boom

A repeating pattern of a photograph of a silicon chip, recolored so that it is multi-colored, in the style of pop art.

Why the AI Policy Debate Should Focus More on the Harness and Protocol Layers

President Donald Trump stands at a podium with the U.S. presidential seal, speaking into a microphone on a stage lit in red and blue. American flags flank both sides, and large signs in front read “Winning the AI Race,” with a backdrop displaying “The Hill & Valley Forum” and “ALL-IN.”

President Donald Trump delivers remarks at the White House AI Summit at Andrew W. Mellon Auditorium in Washington, D.C., Wednesday, July 23, 2025. (

Trump Signs Previously Shelved AI Executive Order

Larry Ellison, executive chairman of Oracle Corp., via video link, and Tony Blair, former UK prime minister, at the World Governments Summit in Dubai, United Arab Emirates, on Wednesday, Feb. 12, 2025. Photographer: Christopher Pike/Bloomberg via Getty Images

The AI ‘Revolution' is Not a People's Revolution

European Commission President Ursula von der Leyen addresses the audience during the conference ‘One Year After the Draghi Report’ at the EU Charlemagne building on September 16, 2025 in Brussels, Belgium. (Photo by Omar Havana/Getty Images)

The Spillover Effects of EU’s Digital Deregulation Agenda

A collage of circuit boards interwoven with aerial photographs of water treatment facilities, agricultural lands, industrial sites and mining machinery.

Rethinking Norm-Setting in Africa’s Digital Present

Fans that are part of a cooling system are seen on the roof of a data center, Monday, April 27, 2026, in Hillsboro, Ore. (AP Photo/Jenny Kane)

The State of Data Center Policy in the United States

Henna Virkkunen, Executive Vice-President for Tech Sovereignty, Security and Democracy of the European Commission. 

EU to Roll Out Tech Sovereignty Agenda to Rival US and China. Will it Work?

Workers continue building the cage for a future UFC fight on the South Lawn in front of the White House, Saturday, May 30 2026, in Washington. (AP Photo/Jose Luis Magana)

How to Rebuild US Science and Technology Policy From the Rubble of Trump 2.0

An air defense unit of the Ukrainian Armed Forces, known as drone hunters, illuminates the sky with a powerful searchlight at night at a position in the Kyiv region of Ukraine, March 24, 2024. Shutterstock

An air defense unit of the Ukrainian Armed Forces, known as drone hunters, illuminates the sky with a powerful searchlight at night at a position in the Kyiv region of Ukraine, March 24, 2024. 

What Middle Powers Risk When AI Goes to War

Courtney C. Radsch, PhD, is a journalist, scholar and human rights advocate who currently directs the Center for Journalism & Liberty at Open Markets Institute and is a non-resident fellow at the Brookings Institution. She serves on the board of Tech Policy Press.

On Wednesday, the UK’s Competition and Markets Authority (CMA) 

 forcing Google to give publishers real control over whether their content powers AI-generated search summaries. Under the 

, publishers can opt out of having their content used in AI Overviews, AI Mode, and broader generative AI services at both directory and page level. Google must also attribute publisher content with clear links in AI-generated results and submit compliance reports every six months. And critically, the CMA adopted one of the key provisions the center I run 

, meaning that Google cannot downrank or penalize publishers in general search results for opting out of the AI features.

These requirements, though they will not be mandated for implementation immediately, are already having ramifications for publishers around the world. Google 

 it will begin testing new controls for a subset of website owners before a global rollout—which means today’s decision has implications for publishers everywhere.

 the CMA to make publisher remedies an early enforcement priority, not something deferred to a later phase when AI markets will have 

 around business models built on the uncompensated extraction of journalistic value. And apparently that argument landed, because publisher protections are among the first conduct requirements being imposed rather than being relegated to a later stage or an afterthought. As I said in our 

, “For too long, dominant digital platforms have been permitted to extract value from journalism while weakening the economic foundations of the organizations that produce it.”

The CMA acted under the UK’s Digital Markets, Competition and Consumers Act 2024, which designated Google with 

 in October 2025, recognizing that it controls over 90% of UK general search queries and other market dynamics that we highlighted in 

 to the consultation. The publisher conduct requirement is the first such requirement imposed on Google, and CMA chief executive Sarah Cardell was explicit that more is coming (hopefully—the CMA has the 

One of the core underlying problems is one we have been documenting for years and is tied to Google’s illegal (in the US at least) monopoly of search and ability to tie this dominance to other products like AI: publishers must allow Google’s crawler to access their content to appear in search results. That same crawled content now powers AI Overviews that reduce traffic back to those publishers—what the CMA aptly calls 

. When AI Overviews appear in a search, the 

, making it impossible to duel digital advertising, monetize audiences through subscriptions or memberships, or cultivate brand awareness, undermining the economic rationale for producing the content Google’s AI depends on.

The timeline is problematic, however, since the main obligations come into force December 3, 2026, with nine months after that for Google to implement. Compliance reports are supposed to land every six months, meaning we may not know whether any of this is working until late 2027. Publishers absorbing losses right now cannot wait that long for evidence that a remedy is functioning. As my CMDG colleague Karina Montoya 

, “meaningful protections will require regulators to ensure these requirements produce real-world outcomes, not merely procedural compliance.”

 by Brazil’s competition authority in April to 

 into Google’s use of journalistic content in Search and Google News rather than accept a recommendation to close the case. CMDG had urged CADE to reopen the inquiry in a 

, alongside allies in the CTRL+J alliance, drawing on evidence from publishers around the world. The authorities recognized that AI Overviews had transformed what was originally a snippets dispute into something considerably more serious, and sent the case back for formal administrative proceedings with expanded scope. The fact that civil society engaged in the process in writing, in person, and across jurisdictions is what is moving these proceedings forward and securing our ability to push back against Google’s dominance.

These developments, alongside similar inquiries in 

 and other jurisdictions, indicate a converging international acknowledgment that Google’s dominance in search and its integration of generative AI cannot be treated as separate from each other, much less separate from the crisis facing journalism and the information ecosystem more broadly. The individuals and organizations producing the 

 cannot be an afterthought in the governance of digital markets.

Independent journalism is essential public infrastructure. If AI systems derive value from news content, the organizations that invest in reporting, fact-finding, and accountability journalism must have meaningful leverage in determining the terms of that relationship. Today is a first step toward that, though the measure of its success will be its enforcement.

Late last week, the United States Department of Justice (DOJ) 

 to Reddit and X (formerly Twitter) demanding identifying information about users who have criticized the Trump administration’s immigration enforcement activities—and directing Reddit and X to send the requested information directly to Immigration and Customs Enforcement (ICE) officials.

These subpoenas are the latest in a persistent campaign by the Trump administration to 

 opposition to and reporting on immigration raids and the federal agents involved, including crowdsourcing and sharing that information online.

Meanwhile, hundreds of people in at least five ICE detention facilities 

 have launched hunger and labor strikes in recent weeks to protest 

 ranging from spoiled food to medical neglect. Outside Delaney Hall, a detention center in Newark, New Jersey, protestors and journalists have been 

 federal agents and state police as community members gather in solidarity with those still on strike inside.

The strikes and related demonstrations are an urgent reminder that the Trump administration continues to target ICE watch efforts for one clear reason: impunity thrives without witnesses.

A fishing expedition enabled by tech companies

The Trump administration’s subpoena campaign to unmask the operators of accounts critical of—or simply reporting on—ICE activities has been ongoing for months.

In early September, roughly three months into escalating DHS operations in Los Angeles, the agency 

 an administrative subpoena to Meta demanding information about multiple users, including the operators of the Instagram account for StopICE.net.

 “hundreds of administrative subpoenas” not only to Meta, but also to Google, Reddit, and Discord demanding identifying information about users behind accounts monitoring ICE activities.

For the most part, these subpoenas have been administrative—meaning companies have no legal obligation to supply the information requested. Still, in many cases targeting individual users, it has been the users themselves who have blocked the release of their personal information by 

Leading tech companies, on the other hand, have openly complied with requests from Trump administration officials to crack down on community ICE watching efforts. In October, at the height of DHS’s chaotic “Midway Blitz” operation in Chicago, Google and Apple removed multiple apps designed to crowdsource community reports about ICE raids from their respective app stores 

 the developer of one targeted app, ICEBlock, that his app violated App Store policies by providing information that might be used to “harm [law enforcement] officers individually or as a group.”

Three days later, on the south side of Chicago, community observer Marimar Martinez was 

 five times by a Border Patrol agent she was filming from her car.

 blocking the federal government from coercing Apple to continue blocking another targeted ICE watch app—Eyes Up—on the grounds that it was violating the app creator’s First Amendment rights. Apple has not relisted either app.

DOJ enters the social media subpoena fray

 sent to Reddit and X are reportedly grand jury subpoenas issued by the US Attorney’s Office for Washington, D.C.—currently led by former FOX News personality Jeanine Pirro.

Attorneys for those targeted by the Reddit and X subpoenas say there is 

 as to what the underlying charges might be.

Under Pirro’s leadership, the office has 

 to convince grand juries to sign off on criminal charges—including the multiple grand juries who refused to indict a man who threw a footlong Subway sub at federal agents during the Trump administration’s federal surge in Washington D.C. last summer while 

 “SHAME!” and calling the agents “racists” and “fascists.” Though Pirro’s office eventually secured an indictment, the government then 

In February, Pirro’s office also tried and 

 to secure a grand jury indictment against multiple Democratic members of Congress—including Senator Mark Kelly (D-Ariz.) and five former military and intelligence officers—who 

 to active-duty military and intelligence officers reminding them that they have the legal right and obligation to refuse illegal orders.

While doggedly pursuing its theory that observing, recording, and criticizing federal agents’ taxpayer-funded activities is a criminal enterprise, the Trump administration continues to 

 the web of sophisticated, invasive surveillance tools at those same agents’ disposal.

 facial recognition apps like MobileFortify from their phones; ICE has 

 access to an “unprecedented” amount of the public’s personal data from other agencies, including the IRS; ICE and Border Patrol are 

 the growing network of automated license plate readers like Flock now blanketing many US cities and towns; and DHS has 

 with surveillance tech companies like Penlink in order to buy mobile phone location data without a warrant.

 about taunting community members—and ICE watch volunteers in particular—with evidence that their activities are being tracked and their personal information is accessible by individual ICE and Border Patrol agents. Lawsuits have been 

 by citizens in Maine and Minnesota describing federal agents “show[ing] up outside the homes of ICE observers” and placing intimidating phone calls alleging that ICE watch volunteers “might get added to a domestic terrorism watchlist” if they continue their work.

While there is no public evidence that any “domestic terrorism watchlist” exists, Trump has 

 efforts to publicize “identifying information” about ICE agents as “violent and terroristic activities.”

Top administration officials have also repeatedly made claims that victims of ICE and Border Patrol violence were actually engaged in acts of “domestic terrorism”—

 two community observers shot and killed by federal agents in Minnesota, Renee Good and Alex Pretti, 

Outside of Delaney Hall, protests in solidarity with striking detainees entered their second week with a curfew in effect for a half-mile radius around the facility.

 demonstrators from gathering close enough to the facility that people trapped inside could hear those gathered outside in support. Dozens of protesters and journalists have been 

 federal agents deliberately targeting them and their cameras.

Local elected officials have repeatedly been denied entry to the facility; Congresswoman LaMonica McIver (D-N.J.), who represents the district where Delaney Hall is located, 

 federal charges for her efforts to gain access to conduct routine oversight last year. Those who have recently been allowed to visit include House Minority Leader Hakeem Jeffries (D-N.Y.), who 

 that what he saw and discussed with detainees “shocked the conscience.”

Delaney Hall is operated by GEO Group, the largest private prison operator in the US. In recent years, the company has also “built a lucrative side business” focused on 

, from phone apps to smart watches, specifically to cash in on immigration enforcement.

GEO Group is no stranger to allegations of abuse and mistreatment; people held in their detention centers have 

 gone on hunger strikes to protest inhumane conditions. The company has faced multiple lawsuits over forcing immigrant detainees to work for just $1 per day, and 

 the allegations made by striking detainees at GEO-owned detention centers 

 that hunger strikes are underway at all.

Despite restricted access for elected officials, attorneys, and families, those trapped inside Delaney Hall 

 is one that broke through. Soto’s wife Gabriela spoke to demonstrators outside Delaney Hall to 

 her husband’s account of conditions inside the facility, where he’d been held for five months. He has since been 

 and reportedly placed in solitary confinement in retaliation for his suspected involvement with the ongoing hunger and labor strike.

In detention centers, in the streets, and on social media, the Trump administration is waging an aggressive campaign against those exposing its worst abuses.

But the message from those still risking retaliation in Delaney Hall and other detention facilities across the country is clear: the costs of looking away are higher than the costs of speaking out.

Hanna Barakat & Archival Images of AI + AIxDESIGN / Better Images of AI / 

 that Anthropic's Claude Mythos Preview could successfully exploit 157 of 898 software vulnerabilities in an AI evaluation benchmark loaded with a data set of known real-world examples. OpenAI's GPT-5.5 was not far behind, successfully exploiting 120 of 898 known vulnerabilities. This has led to predictions of “bugmaggedon,” the idea that when released to the public, Mythos and similarly powerful models will hand cyberattackers the ability to compromise the digital infrastructure underlying all of modern life faster than defenders can protect it.

 last month exploited a vulnerability in widely used educational software called Canvas to steal usernames, email addresses, enrollment information and private messages between students and teachers, disrupting classrooms for 30 million users at over 8,000 colleges, universities and K-12 schools worldwide. The educational technology sector already 

 the largest financial institutions, the federal government, enterprise software companies and other well-resourced players who are 

 these kinds of vulnerabilities before bugmageddon materializes. So how do we help to prepare our schools and the many important organizations that form the backbone of our communities?

The capabilities of frontier AI models like Mythos will disproportionately affect organizations that are already least prepared to protect themselves or recover from cyberattacks. “Patch Tuesday” is not a realistic expectation for the small dentist’s office, the food bank, the volunteer fire department, the public school, or the rural electric co-op and water utility. They already operate in a degraded security environment because they frequently lack the budget and/or in-house expertise to implement robust cybersecurity programs. Still, we ask them to fend for themselves against sophisticated cybercriminal gangs and nation-state adversaries who will soon be equipped with even more powerful AI-enabled cyber offense capabilities.

Free tools, tokens and cybersecurity software alone are insufficient to solve a problem that is more about humans than it is about technology. The idea that millions of cyber-poor community organizations will stop snoozing software updates or have the time and resources to train and deploy their own cyber-defense agents ignores the reality on the ground. Hands-on support is imperative to improve cybersecurity outcomes for community-serving organizations. A 

 by the Cyber Readiness Institute found that a water utility working with a human coach was 3 times more likely to complete a cyber resilience program than a utility that tackled the program alone. State governments have 

 and have created teams of cybersecurity volunteers to help cities and schools during cyber emergencies.

Trusted human networks are still the best way to deliver cybersecurity help. In an environment where bugs are found faster than humans can patch them and we rely on AI for cyber defense, access to hands-on cybersecurity expertise is essential.

Luckily, these local support networks exist, and they need to be expanded and scaled. Consider just some of the public service-minded groups already working in communities to help. Members of the 

 —including the Consortium of Cybersecurity Clinics; 

, Maryland and Ohio; DEFCON Franklin and the 

 — have deployed thousands of knowledgeable cyber volunteers.

 (RSOCs) in Arizona, Texas, Oregon and several other states provide ongoing cybersecurity support pro bono to public agencies, critical infrastructure and other community organizations. The 

 program in Indiana has reached K-12 school districts, public libraries and municipal governments in the majority of the state’s counties. 

 are stepping up to fill gaps in managed security services for community organizations unable to afford corporate consulting rates. The 

 build on a long tradition of federally funded technical assistance for our water systems. And research universities, including Berkeley and 

 are bridging research and practice to better understand and replicate what works best.

The task now is to turbo-charge these human-centered initiatives to meet the new reality of AI’s capabilities. Anthropic calls its pre-release regime 

, in reference to the glasswing butterfly, which can hide in plain sight. 

Let's call this effort Project Kaleidoscope — the swarm of butterflies that we need to defend our communities. Here is a roadmap for how to put it into action:

Scale up cyber volunteering through existing platforms such as university-based cybersecurity clinics, state-based civilian cyber corps, and corporate volunteering.

Create and fund shared service hubs, such as RSOCs and managed security services providers (MSSPs), that could provide sustained cybersecurity support to community organizations at below-market cost. These hubs could partner with the frontier AI labs for early access to defensive models and deploy models like the cyber circuit riders to reach their constituencies.

Develop regional ecosystems and “connective tissue” that link volunteers, shared services and other government-provided resources into long-term support systems.

Embed cyber knowledge in communities, so organizations strengthen their understanding of cybersecurity and their role in protecting themselves.

Government, academia, industry and civil society all have a role to play, and we can act regardless of whether or how soon federal AI regulation arrives. Cyberattacks on community-serving organizations and critical infrastructure put everyone at risk and reveal our interdependencies. Hospitals need water. Small businesses need childcare. Utilities need city governments, and city governments need nonprofits

A successful cyberattack on one organization has ripple effects across an entire community, and we can’t afford to wait for bugmaggedon to occur. It’s time to include community organizations in our responses and invest in what really works on the ground: people who can help.

This three-part series is published in partnership with the AI Now Institute.

Europe, Spring 2026. Germany’s AI translation darling DeepL 

 with Amazon Web Services. French Mistral 

 their models in partnership with Nvidia. Swedish Lovable 

 a jump scare as Anthropic moves in to eat its lunch. Meanwhile, across Europe, companies who submitted bids to build the continent's ‘Gigafactories’ — large-scale computing infrastructure designed to help Europe train the latest generation of frontier AI models — are 

For sovereignty-minded EU politicians, this landscape might cause distress. Massive amounts of political energy are being spent to break free from US Big Tech: by deregulating, providing capital, aspiring to build sovereign computing clusters, and nurturing ambitious tech startups. Yet, the initiatives keep being frustrated by the gravity wells of the global tech market that bend the interests of private actors in their direction.

 the growing political effort to build "Sovereign AI” in Europe. A key theme throughout this work has been that Europe’s AI industrial policy rests on an implicit, largely unexamined assumption: that European AI companies, their funders, and their market incentives are naturally aligned with European strategic interests. Simply put, the more European AI companies there are, the better for Europe.

The reality is much messier. What looks like a European AI ecosystem is a confluence of interests pulling in different directions: VC timelines that reward fast exits over long-term independence, market structures that funnel revenue upstream to US providers, infrastructure dependencies that deepen with every investment, and playbooks borrowed wholesale from Silicon Valley.

The policy problem runs even deeper. This week, the European Commission presented the 

European Technological Sovereignty Package

, a set of measures to strengthen Europe's capacity in semiconductors, AI, cloud and open source. This is more than a framing exercise. With the proposed 

, the package contains two concrete legislative proposals that are deeply revealing about the unspoken assumptions that drive Europe’s strategy on AI: about where value is created and captured in AI supply chains, about AI's future trajectory, and about whether sovereignty can be achieved without engaging in market shaping.

Simply put, Europe’s strategy on AI has focussed on two sides of the market: boosting demand (via initiatives like the Apply AI Strategy, aimed at pushing AI adoption), and — primarily — increasing supply of inputs (via the Cloud and AI Development Act, expanding on compute and data center capacity and the proposed Cloud Act 2.0, aimed at increasing Europe’s chips supply). What was initially framed under the umbrella of competitiveness, now includes added measures aimed at ensuring “sovereignty,” with the laudable goal of progressing towards a full European technology stack.

As we will argue throughout this series, these industrial policy proposals rest on an incomplete, two-dimensional view of the European AI market. In reality, this market is 

 with the ecosystems of dominant US players, in ways that boosting both supply and demand alone cannot disentangle. As a result, interventions that seek to secure sovereignty in AI risk leading to much more entrenched dependence.

Most EU tech policy commentary operates top-down: it starts from strategy documents and asks whether targets are being met. This series works bottom-up. It looks at what the European AI market actually looks like (who supplies whom, where revenue flows, what dependencies are baked in) and tests the assumptions that policy documents leave implicit. Where evidence is incomplete or trajectories genuinely uncertain, we will say so openly rather than defaulting to confident prescriptions.

This mini-series consists of three parts: The first piece will examine Europe's AI application layer; how startups like Lovable function as a distribution channel for US models, and what that means for the value Europe actually captures. The second will trace why even Europe's strongest independent AI companies, like DeepL, end up dependent on US hyperscaler infrastructure. The third will pull the threads together: the common structural dynamic, the strategic choices Europe is avoiding, and the honest uncertainties about where AI is heading. These pieces will be published over the coming weeks in Tech Policy Press and on 

the EU AI Industrial Policy Monitor Newsletter

The obvious risk is that the widespread adoption of AI further entrenches Europe’s dependence on a handful of US firms. The less obvious risk — and the one this series examines — is that even where Europe builds its own AI capabilities, the economic value still flows elsewhere. Whether and how much to build is itself a political question this series doesn't presume to answer. But whatever Europe decides, an AI industrial policy strategy that cannot decipher its priors — particularly around where and how it will complete — will not be successful. It is a wish list that avoids the scrutiny that comes from clearly naming tradeoffs and hard choices.


The World Is Already Resisting AI. Now, There is a List to Prove It.

Brazil's 2026 Elections Are Its First Real Stress Test for AI Regulation

Surveillance Technology is Silencing Journalists in Kashmir

The Meta logo on a smartphone in front of an image of CEO Mark Zuckerberg.

Meta’s Worker Surveillance Tests EU Rules on AI and Labor

Tech Policy Press - Technology and Democracy

Tech Policy Press is tracking laws and regulations, along with government investigations and litigation, that will shape the rules and accountability for tech companies.

Latest

Podcast

Recent

Our content delivered to your inbox

Thank you!

Around the globe