The AI sovereignty problem, from Brussels to Bengaluru

Good morning!

When the fictional near-future scenario Europe 2031 was published in June, it set off an immediate debate in Brussels and beyond. Co-authored by a group of AI researchers, its purpose was to provoke. The scenario attempts to demonstrate that Europe risks economic and political irrelevance if it fails to compete at the frontier of AI development. Within days of its publication, the United States government ordered Anthropic to restrict access to its most advanced AI model for non-US users, an event the scenario had projected for 2029.

In this week’s Tech Policy Press podcast, Deputy Editor Ramsha Jahangir is joined by one of the co-authors of Europe 2031, Maximilian Negele, and the AI Now Institute’s Frederike Kaltheuner, whose three-part series for Tech Policy Press argues that Europe's dependency problem runs deeper than frontier model access, to debate what Europe is actually getting wrong, and what it would take to change course. Listen to the podcast here.

Questions about the geopolitics of AI and digital sovereignty surfaced repeatedly across Tech Policy Press this week.

• Europe cannot build a chokepoint on someone else's supply chain. Wim Casteels, a lecturer and researcher at AP University College of Applied Sciences and Arts in Antwerp, Belgium, on why ASML's semiconductor monopoly doesn't give Europe strategic control.
• Two AI superpowers, shared risks, no shared rules. Pedro Kritski, an engineer from the Federal University of Technology–Paraná (UTFPR), and Virgilio Almeida, a faculty associate at Harvard’s Berkman Klein Center for Internet and Society and professor emeritus of computer science at the Federal University of Minas Gerais (UFMG), Brazil, argue that fear of autonomous weapons should mobilize democratic debate. It "cannot be left solely to governments, armed forces, or technology companies"—it must involve civil society before it's too late.
• The UN held its AI for Good Global Summit following the Global Dialogue on Artificial Intelligence Governance in Geneva last week, journalist Chris Stokel-Walker reports. Heads of state and UN leaders were joined by executives from companies such as Nvidia, Amazon, Microsoft and ZTE.
• Relatedly, researchers at Umeå University's AI Policy Lab argue the UN Scientific Panel on AI's first report leaves its independence unresolved—from both industry and states—and propose five procedures and disclosures to fix it.
• The European Commission has joined W Social, a closed-source European platform, just weeks after publishing its Open Source Strategy. Megan Kirkwood, a researcher and writer specializing in issues related to competition and antitrust, asks whether the move contradicts the Commission's own goals.

Spotlight on the US Supreme Court

• Tech Policy Press fellow Jake Laperruque and privacy and cybersecurity attorney Katelyn Ringrose call the 6-3 decision in Chatrie v. United States one of the US Supreme Court's most important Fourth Amendment rulings of the 21st century, and outline its key impacts and open questions for digital privacy. “Even as Chatrie bolsters Fourth Amendment protections for digital data, the government will likely continue to rely on other arguments, such as ICE’s spurious claim that purchasing location data tied to Mobile Advertising IDs (which can be de-anonymized) rather than phone numbers nullifies any Fourth Amendment concerns,” they write.
• For centuries, courts and lawmakers have been the primary regulators of law enforcement investigations, placing legal limits on how police can collect information. Now, companies like Google are becoming just as influential, writes Texas A&M University School of Law professor Hannah Bloch-Wehba, reflecting on the outcome in Chatrie.
• "Do we really want a president to be powerless to implement the protections for consumers, workers and the environment on which he or she ran?" asks Georgetown's Mark MacCarthy—arguing the Supreme Court result in Trump v. Slaughter "might even have some upsides for a new progressive administration."

Protection or restriction?: Children’s online safety

• Tech Policy Press contributing editor Mark Scott sums up where the EU stands ahead of its child safety panel's July 13 recommendations — and whether European Commission President Ursula von der Leyen will push for an Australia-style teen social media ban.
• Europe appears to believe it has two options for protecting childrenonline: scan their messages or ban them from platforms. Tanya Becheva, founder of independent Danish nonprofit Custorian.org, argues there is a third way, and Europe has been too busy fighting over the first two to build it. “That is why we need a standard: an open, auditable definition of what a safe-for-children service actually does, control by control, that an independent third party can certify against — the way we certify food hygiene, payment security, or building fire safety,” she writes.
• Bans regulate users, and taxes on ad-revenue target the wrong variable; a tax on harmful amplification would reach the producer's incentives instead, writes Mihaela Popa-Wyatt, senior lecturer in Philosophy at the University of Manchester.

Surveillance and sovereignty in India

• Discriminatory instincts, encoded into flawed databases, handed back to police as algorithmic confirmation. Asvatha Babu, an independent researcher based in Chennai, India, maps the injustice baked into one of the world's most expansive surveillance systems—a system most Indians don't know exists. “Unfortunately, the line that separates a good citizen from a bad one is often a politically drawn one that can be easily shifted to serve the prerogative of those in power,” she writes.
• Digital sovereignty has become less about eliminating dependency than about choosing among competing foreign partners. Meg Kitamura, a doctoral researcher at the Center for Critical Computational Studies at Goethe University Frankfurt, on what India's Aadhaar reveals about that shift. “While digital sovereignty is increasingly being mobilized as a solution to the ailments of Big Tech authoritarianism, digital sovereignty in the form of self-determination and technological independence over a select few geopolitical rivals must not be mistaken for a pathway to democratic accountability,” writes Kitamura.

AI oversight and governance gaps

• Responding to Dario Amodei's call for industry self-governance, Oversight Board members Kenji Yoshino and Ronaldo Lemos lay out three requirements for credible AI oversight: independent overseers, external standards, and transparency.
• Danai Nhando, a global advisor at the Atlantic Council's Freedom and Prosperity Center and AI, examines how AI adoption is outpacing governance across the humanitarian and nonprofit sectors, and what closing that gap requires. When AI governance fails in aid work, she writes, "the people who pay the price are the ones already fighting for their lives"—with no say in any of it.
• Sumit Sharma, an independent economist, advocate, and tech policy expert, argues antitrust enforcement against Big Tech–AI ownership crossholdings could prevent a popping AI bubble from cascading across the economy, helping to restore the competitive incentives that crossholdings have dampened. “We need antitrust enforcement to ensure that AI firms compete independently and vigorously, without the contractual restraints of cloud-for-equity, without overlapping boards, and without minority stakes that quietly align incentives,” he writes.
• By July 1, US states enacted 109 AI laws and 28 data center laws so far in 2026, writes NYU Center on Technology Policy director Scott Babwah Brennen—despite federal efforts to stop or limit state AI legislation. He offers eight observations on the state of state AI laws.
• The real AI hazard isn't just misalignment, argue Jules Polonetsky, CEO of the Future of Privacy Forum, and Omer Tene, legal practitioner and thought leader on privacy, cybersecurity, AI and law, and technology. It's Babel: teaching ourselves to treat "speech as output, judgment as computation and persons as data." The idol diminishes us before it ever comes alive.

Accounting for AI’s social and environmental harms

• Based on 10 months of fieldwork, Hannah Lipstein, now a research associate at the Distributed AI Research Institute (DAIR), shares five takeaways from "Greening AI in the public sector," a new procurement handbook produced with Data & Society and the GovAI Coalition for government agencies and citizens weighing AI's environmental impacts.
• Work in the AI economy is deliberately organized through an "outsourcing maze," writes the Weizenbaum Institute’s Sana Ahmad, enabling global capital to extract value from workers via layered intermediaries.

The information battlefield

• Following multiple instances of social media-fueled violence, Demos Digital interim director Hannah Perry argues the next UK government should move the Defending Democracy Taskforce into Downing Street and publish a new strategy for information ecosystems resilience. “Agentic AI is paving a future where citizens no longer need to enter shared online spaces at all, introducing a new set of challenges to the equation. Epistemic security is at risk,” she writes.
• Jordyn Abrams of GWU's Program on Extremism examines how apocalyptic AI rhetoric from tech executives may be fueling a rise in anti-tech violence—and argues for sober, fact-based discourse instead.
• Policymakers treat critical infrastructure as physical: power grids, communications networks, transportation systems, and satellites. In the age of generative AI, writes Rachel A. George, a lecturer in international relations at Stanford and a Legal Innovation & Policy Fellow at Stanford Law School, another kind deserves equal attention: the institutions that determine what governments believe to be true.
• Major platforms sold political ads in Cyprus despite saying they had banned the practice, according to new research. Tech Policy Press fellow Liz Carolan on the gaps in EU rules ahead of major European elections. “The researcher found that Meta platforms allowed at least 2,457 ads to pass internal checks and be pushed to voters,” she reports.
• Russian networks have already flooded the web with content designed to corrupt AI training data. Gerald Mako, a research affiliate of the Cambridge Central Asia Forum at Cambridge University, writes that agentic AI makes this threat far worse — and Brussels is still playing catch-up.

Global digital policy roundup

• Digital Policy Alert's Maria Buza and Aishwarya Vaithyanathan round up June 2026 developments in global digital policy across content moderation, AI, competition, and data governance across the G20. From new legislation to fines and enforcement, check out what's new. These roundups are an incredible resource; I always find something I missed.

What we’re watching

On Friday, the European Commission issued a preliminary finding that Meta's Instagram and Facebook violate the Digital Services Act due to addictive design features that breach consumer protection laws. Meta faces potential fines and is required to make design changes to address the addictive nature of its platforms. We’ll host coverage of the finding next week, and we invite your perspectives.

To those fans mourning in Norway, Switzerland and beyond, I send my condolences. But it is remarkable that the four top-ranked teams are competing in the World Cup semi-finals; something to look forward to for the week ahead!

-Justin