A brightly coloured mural which can be viewed in any direction. It has several scenes within it: people in front of computers seeming stressed, a number of faces overlaid over each other, squashed emojis, miners digging in front of a huge mountain representing mineral resources, a hand holding a lump of coal or carbon, hands manipulating stock charts and error messages, as well as some women performing tasks on computers, men in suits around a table, someone in a data centre, big hands controlling the scenes and holding a phone, people in a production line. Motifs such as network diagrams and melting emojis are placed throughout the busy vignettes.


Making the Case for Actionable Prosocial Design Research

Can the Pall Mall Process Curb Spyware Abuse?

United Nations Secretary-General António Guterres addresses the opening ceremony of the 2026 World AI Conference and High-Level Meeting on Global AI Governance on July 17, 2026 in Shanghai, China. Themed "AI Partnership for a Brighter Future," the event included more than 140 forums, bringing together 1,400 guests from home and abroad. (Photo by VCG/VCG via AP)

United Nations Secretary-General António Guterres addresses the opening ceremony of the 2026 

World AI Conference and High-Level Meeting on Global AI Governance

 on July 17, 2026 in Shanghai, China. Themed "AI Partnership for a Brighter Future," the event included more than 140 forums, bringing together 1,400 guests from home and abroad. (Photo by VCG/VCG via AP)

UN Scientific Panel Warns of AI Concentration, But UN’s Vision Entrenches It

Flock Safety camera at the intersection of West Sylvania Avenue and Homeville Circle in Toledo, Ohio. The ubiquity of Flock surveillance systems in the United States raises questions about privacy and data security. (Photo by Stephen Zenner / SOPA Images/Sipa USA)(Sipa via AP Images)

Flock Safety camera at the intersection of West Sylvania Avenue and Homeville Circle in Toledo, Ohio. The ubiquity of Flock surveillance systems in the United States raises questions about privacy and data security. (Photo by Stephen Zenner / Sipa via AP Images)

Researchers Detail How AI Systems Can Enable Authoritarianism

Portland, OR, USA - Sep 1, 2024: ChatGPT, Gemini, Microsoft Copilot, Claude, and Perplexity app icons are seen on a Google Pixel smartphone. AI competition concepts.

ChatGPT Didn't Break the AI Act. It Showed Why Adaptive Regulation Matters

Social media puppeteering the public through paid political influence — a hand emerging from a smartphone controls a wooden marionette by red strings.

How to Close the Federal Election Commission’s Dark Money Loophole

European Union flag displayed on a screen and AliExpress, Temu and Shein icons displayed on a phone screen are seen in this illustration photo taken in Krakow, Poland on July 1, 2026. (Photo by Jakub Porzycki/NurPhoto via AP)

(Photo by Jakub Porzycki/NurPhoto via AP)

EU Turns to Online Marketplaces After Record AliExpress Fine

BIDDEFORD, MAINE - JULY 15: A person observes a makeshift memorial for Joan Sebastian Durán Guerrero, a 26-year-old Colombian immigrant who was fatally shot by US Immigration and Customs Enforcement (ICE) agents, on July 15, 2026 in Biddeford, Maine. Guerrero's death comes less than a week after federal immigration agents shot and killed 52-year-old Mexican national Lorenzo Salgado Araujo in Texas. (Photo by Ryan Murphy/Getty Images)

ICE and the Architecture of Authoritarianism

Spain fans gather for a rally on the eve of the World Cup final soccer match in New York's Times Square, Saturday, July 18, 2026 (AP Photo/Yuki Iwamura)

The Silent World Cup Winners are the Companies Normalizing Surveillance

Miami, June 24, 2026: WORLD CUP, BRAZIL vs. SCOTLAND. General view of the VAR screen at the Miami stadium for the match between Brazil and Scotland in the 2026 World Cup. Photo: Rodolfo Buhrer/AGIF

Soccer's Tech Revolution Has a Labor Chain

A megaphone with the words vote vote vote coming out of it

Chatbot Voting Advice Could Help and Harm Voters. Let’s Regulate Accordingly.

Thumbs-down hand with a one-star rating painted on the forearm, set against a teal background — representing a negative review or low rating.

What Review Bombing Tells Us About the Future of Dissent in Everyday Life

March 10, 2026: The London Book Fair (LBF) at Kensington Olympia. Names of authors in an otherwise blank book titled Don't Steal This Book protesting against AI companies taking authors’ works. (Matthew Chattle/Future Publishing via Getty Images)

March 10, 2026: The London Book Fair (LBF) at Kensington Olympia. Names of authors in an otherwise blank book titled 

 protesting against AI companies taking authors’ works. (Matthew Chattle/Future Publishing via Getty Images)

Copyright Law Wasn’t Built for the AI Era. We Need ‘Learnright.'

We Need Political Philosophy for Mental Health Chatbots

The United Nations General Assembly Hall in New York City. Shutterstock

The United Nations General Assembly Hall in New York City. 

Blocking Experts Will Weaken the UN's New Cyber Rules

A humanoid robot developed by Geneva-based technology company RB Labs runs through the exhibition aisles during a showcase at the AI for Good Global Summit, a United Nations flagship event aimed at shaping the future of artificial intelligence, in Geneva on July 7, 2026. (Photo by Fabrice COFFRINI / AFP via Getty Images)

How the UN’s Scientific Panel Erases Human Responsibility for AI

Collage-style illustration on a blue background showing a laptop displaying "AI" merged with a large eyeball, a human hand typing on the keyboard, a gear, a magnifying glass, and an orbiting ring, evoking themes of AI agents and automated search.

We Can’t Monitor AI Agents at Scale. Here’s What It Will Take.

This photograph shows Robert the Robot, an advanced conversational humanoid developed by Geneva-based technology company RB Labs, at the AI for Good Global Summit, a United Nations flagship event aimed at shaping the future of artificial intelligence, in Geneva on July 7, 2026. (Photo by Fabrice COFFRINI / AFP via Getty Images)

The Internet Had a North Star. The UN’s Global Dialogue Made Clear AI Doesn't.

United Nations Secretary-General António Guterres speaks at the opening of the first session of the Global Dialogue on AI Governance in Geneva, Switzerland, July 6, 2026. (Photo by Lian Yi/Xinhua via Getty Images)

After Geneva, AI Governance Must Confront the Trust Deficit

Two workers running inside a scroll wheel embedded in a computer mouse, controlled by a giant capitalist hand.

AI Lays Bare the Authoritarianism of Modern Work. Time to Rethink Education.

, the European Commission held that the "addictive design" of Instagram and Facebook 

, citing harms to users' physical and mental wellbeing. The Commission says Meta should “implement design changes” to both platforms. Such changes may be overdue, but regulating harmful design is only half the equation — attention must also go toward how platforms can build prosocial design, and toward the research that shows them how.

Platforms, universities and organizations in the tech accountability field can — and should — do more to ensure that research on prosocial design informs how healthy digital spaces get built.

Academics and other researchers outside of industry produce hundreds of studies a year to investigate and inform the use of prosocial design: the design features and practices that foster healthy interactions in digital spaces. Yet that research all too infrequently influences practice. Often this is due to familiar barriers, like prosocial design being treated as a secondary 'nice to have,' but tech professionals raise a more specific concern: that much of the independent research that is produced is either not attuned to the systems and constraints of tech platforms, or doesn't address tech professionals' questions and challenges. In other words, it is not “actionable” research that has the potential to inform practice. This isn't to say all research needs to be actionable. Fundamental research has its own vital role, but for research that aims to improve how digital spaces are built, closing the gap between insight and implementation matters.

The case for closing those gaps is built around four ideas: 1) that actionable, prosocial design research is genuinely valuable because it can shape what gets built; 2) that the builders of digital spaces can't be relied on solely to produce it; 3) that independent researchers can fill the void but today lack the pathways and sufficient incentives to do so; and 4) that there are real ways to change that.

Independent research is critical because it is often better equipped to identify useful insights. For example, practitioners participating in the writing of our 

, “Connecting Researchers & Practitioners to Catalyze Actionable Research,” noted how independent, actionable research can “signal avenues of approach that may be more productive than the ones [companies] are already pursuing,” sparking ideas and providing empirical evidence to justify shifts in roadmaps. 

Additionally, practitioners noted how decision makers inside of companies can sometimes fail to see beyond their own business objectives and that synthesized findings from external research presented as “digestible information” can lead directly to “positive outcomes for users” and, in turn, the business itself.

Independent prosocial design research can also support advocacy groups and prosocial-minded professionals working within companies by helping them build the case for the adoption of prosocial design—as well as creators building alternative platforms with prosocial goals at the foundation.

One research product that several practitioners who participated in the writing of our report agreed served as a useful model was “Shouting into the Void,” 

 co-produced by Meedan and PEN America. In it, the authors investigate user reporting experiences across social media and gaming platforms. Leveraging data from interviews and focus groups, they document specific gaps in current reporting systems and describe users’ experiences of reporting. They conclude with both core platform-level and feature-level recommendations. Practitioners found this whitepaper particularly valuable in its ability to both raise awareness of a user need, grounded in research, while at the same time pointing toward possible solutions.

Industry can't be the only source of answers

Platforms can not and should not be the only ones producing research on how to build healthier digital spaces. Prosocial design research should be a public resource, but if and when it’s conducted internally by platforms, it’s rarely shared externally.

Perhaps the greatest impediment lies in the risks to industry of publicly –- or even privately — engaging in research related to prosocial design. These include both brand image as well as legal risks. Publicly working on or being associated with research aimed at creating healthier environments can draw unwanted attention to platforms by implicitly acknowledging that they are not healthy enough, if not unhealthy. The risks can even be legal if, for example, it is shown that platforms were aware of those harms.

Many of the platforms oriented toward creating healthy online environments may be more readily willing to conduct and share research and could fill the gap, but they are often smaller and/or nonprofit ventures that have the fewest resources, even though they may be the most willing to apply lessons learned from prosocial design research (for example, Vermont’s 

While we do not give platforms a pass and incentives or regulation could force the production and disclosure of prosocial design research, the current reality of the disincentive structure described above means we can’t depend on industry to produce insights on how to build for healthier outcomes.

None of this relieves platforms of their responsibilities. But given the legal risks, and the resource constraints, independent researchers may be better placed to produce research that is generalizable across platforms and fills a gap that industry, for structural reasons, cannot fill on its own.

Independent researchers can help fill the gap, but lack the incentives

Sufficient incentives must be in place to encourage researchers to generate practice-informed research. While cross-sector research collaborations would go far to foster actionable research, 

 discuss how mismatches in incentives and resources across the academic, private, and nonprofit sectors impede such collaborations. 

 argue that academic institutions and researchers have de-emphasized civic missions, focusing on ‘publications over the public good.’ At a deeper level, practitioners and researchers may distrust each others' intentions in entering a collaboration and be skeptical of their counterparts' capacity to produce useful, relevant and rigorous insights 

 note that the pace of technology outstrips that of research, resulting in traditional research timelines becoming a “bottleneck for interventions” to address harms before they proliferate and creating a challenge for "temporal validity" (

Those barriers may explain in part why, in a review of 1,288 academic research papers about topics relevant to T&S collected by the Trust & Safety Foundation, 

 found only 5% were collaborations with moderators who help manage online communities, and a mere 3% were collaborations with platform partners. In our report, we make suggestions for how to catalyze independent actionable research and why we think it matters.

Say you're bought into the idea that actionable research matters. Great. How do we make it happen? 

, we outline recommendations for academics and industry to make research more actionable, which we summarize below:

For academic institutions, we suggest rewarding actionable, practice-informed research in faculty hires, reviews and promotions. We also suggest facilitating data-use and other agreements that are necessary for industry partnerships, which can otherwise slow down projects or get stuck in legal departments. Flexibility in partnerships, for example, allowing industry funding to be structured as a gift or as independent contractor funding, can also help.

Industry can help by publicly signaling when independent research informs their work and improves their ability to build healthy spaces so that academics can demonstrate that their research made a difference. Industry can also create titles, such as “research advisor,” for researchers they reach out to for advice or insight on prosocial design, and they can create more internships for social science PhD students. They can also help fund actionable research, but to avoid concerns about “

,” funding needs to safeguard independence and transparency, and should represent a 

. Lastly, industry can encourage staff to get involved by attending conferences where they can make connections with social scientists, or partner on projects with universities.

Civil society organizations, funders and individuals don't have to wait for academia or industry to make the first move. In our report, we propose several projects that go a long way toward reducing the barriers and incentivizing researchers and tech professionals to connect and catalyze actionable research including: surveying practitioners to identify pressing questions, creating a platform for practitioner feedback, co-creating research agendas and hosting researcher-practitioner convenings to workshop design solutions to common challenges, as we did with our recent 

Regulators can require platforms to stop building for harm, but no enforcement action can tell them what to build instead. As the Council on Technology and Social Cohesion’s 

Blueprint on Prosocial Tech Design Governance

 argues, direction has to come from civil society and research, and it will only shape what gets built if researchers, platforms, and funders get together to close the gap between insight and implementation. 

 goes deeper on what makes research actionable, the barriers that make this challenging, examples of research practitioners have found useful, and a fuller set of recommendations for how to get there.

Vas Panagiotopoulos is a fellow at Tech Policy Press.

In early 2024, France and the United Kingdom 

 the Pall Mall Process (PMP), an international initiative that brought together governments, industry, civil society, and academia to tackle “the proliferation and irresponsible use of commercial cyber intrusion capabilities (CCICs),” an umbrella term referring to spyware, software vulnerabilities, hackers-for-hire, and related technologies.

 of spyware abuse, coupled with a growing global demand for such technologies. According to UK intelligence, approximately 100 countries 

 CCICs, highlighting the declining barriers to entry for states seeking these capabilities.

In April 2025, a non-binding Code of Practice for States, now signed by 27 countries, 

, setting out commitments for government action in relation to the development, facilitation, purchase, and use of CCICs. A Code of Practice for Industry is currently being developed and is 

 to be finalized at the Paris Peace Forum in November.

To this end, France and the UK launched a 

 that received 68 responses, including 27 from private industry, among them companies based in the United States, Europe, and the Middle East, according to diplomatic sources.

Among other findings, the consultation highlighted the central role of customer vetting, supply chain due diligence, and transparent business information (e.g. beneficial ownership). However, these practices were implemented unevenly across the market, reflecting differences in companies' size, maturity, and commitment to such safeguards, according to a consultation summary report seen by Tech Policy Press.

PMP has played an important role in making governments recognize the risks that CCICs pose to human rights and national security, particularly by adopting an effective, results-oriented approach.

“In the past, we have seen a number […] of cyber diplomacy initiatives that gathered a large number of states but brought few results,” says Alexandra Paulus, a senior associate on international security at the German Institute for International and Security Affairs (SWP).

“PMP is decidedly not taking that approach. Instead, its initiators seek to bring together a small group of like-minded states, focused on a narrowly scoped problem, to make progress in that field,” she explains.

Such countries include those with significant influence over the global CCIC market (the United States); countries that host spyware vendors, or parts of their multi-jurisdictional corporate structures (e.g. Italy, Luxembourg, and Ireland); and countries where significant abuses involving this technology have occurred (e.g. Poland, Hungary, and Greece).

“The drafting process of the Code of Practice for States sparked a series of government debates about the use of CCICs that had not taken place before,” observes Paulus.

Now, by drafting a code of practice also for businesses, Alexandra Paulus hopes that states will meaningfully engage with industry, too. “Both [codes] can inspire national policies on CCICs,” she says.

Will the PMP contribute to spyware normalization?

Yet, the PMP has also attracted significant criticism. A key point of condemnation from legal experts and civil society is that the PMP – whose scope, it should be noted, does not extend to government-developed CCICs – appears to form part of a broader process of normalizing spyware use.

“Pall Mall moves us from a denialism of use to an embrace of use. And in fact, in many ways, we can see it as a framework that encourages, recognizes, and anticipates further spyware use,” says Fionnuala Ní Aoláin, a former special rapporteur on the promotion and protection of human rights and fundamental freedoms while countering terrorism for the UN Human Rights Council and a professor at the University of Minnesota, who recently 

 a legal and political assessment of the PMP.

Ní Aoláin observes that the PMP’s Code of Practice for States fails to articulate, with sufficient specificity, what states' obligations are in relation to this technology, particularly given its well-documented risks and the requirements of international human rights law.

Similarly, the PMP's framing around responsible use is problematic, as is its failure to define the relationship between domestic law and international frameworks, particularly where these don’t align.

“The Code of Practice for States specifies several criteria for assessing responsible use like domestic law, international law, respect for human rights and fundamental freedoms, and the UN framework for responsible state behavior. However, there is much debate about how each of these should be operationalized and what they mean in practice,” explains SWP’s Alexandra Paulus.

Another point of concern is the nature of civil society's engagement in the PMP, given that activists are often the targets of spyware attacks. “It is ad hoc and unsystematic, and it always feels like civil society is brought in at the last minute,” observes Ní Aoláin.

She argues that the process surrounding the Code of Practice for Industry “is moving very quickly without substantive engagement from human rights experts or civil society.”

Diplomatic sources, however, stress that the PMP's accelerated timeline reflects the “high international demand to act swiftly,” resulting in tight timeframes for all stakeholders.

Finally, two additional points of criticism is that the PMP, in its current form, does not address the need for ‘hard’ regulation to prevent the acquisition of CCICs by non-state groups; as well as concerns that without a robust exclusion capacity the PMP 

 by companies such as NSO Group – whose association with systematic human rights violations is well documented – as a means of rehabilitation and a way to mitigate the impact of sanctions imposed against them. Diplomatic sources maintain that the PMP will not involve consultations with companies or individuals that are sanctioned, subject to restrictions for irresponsible conduct, or facing legal proceedings in participating states' jurisdictions.

Next steps: industry code, implementation and expansion

In the next phase of the PMP, alongside the adoption of the Code of Practice for Industry, France and the UK are seeking to expand global participation through international outreach. Events have already been held in the Western Balkans, Southeast Asia, and Latin America, with a workshop planned for Africa in autumn 2026.

Yet key states where commercial providers of CCICs are based – including Israel and India – are still not participating. This is unsurprising to some extent, but “the absence of countries like Spain, for example, is striking,” says SWP’s Alexandra Paulus.

Securing a commitment from all EU member states to join the Pall Mall Process would send a strong signal, diplomatic sources stress.

France and the UK are also supporting the implementation of the Code of Practice for States through quarterly workshops for signatory governments aimed at strengthening national governance and sharing best practices. They also intend to monitor progress and measure the impact of the PMP's recommendations.

It is worth noting, however, that the Code of Practice for States does not include any accountability mechanisms for participating states that fail to meet their commitments. “Without one, the document becomes a mere declaration of intent that will hardly have political consequences,” says SWP’s Alexandra Paulus.

“But […] for many governments, particularly those with less well-resourced militaries and intelligence agencies, CCICs are a capability that they deem important to national security, and I consider it unlikely that they will voluntarily submit to any accountability mechanism.”

 on July 17, 2026 in Shanghai, China. Themed "AI Partnership for a Brighter Future," the event included more than 140 forums, bringing together 1,400 guests from home and abroad. (Photo by VCG/VCG via AP)


The United Nations Independent International Scientific Panel on AI (IISPAI) is the latest in a growing series of UN initiatives aimed at shaping global AI governance, following efforts such as the 2023 

High-Level Advisory Body on Artificial Intelligence

 and the negotiations leading to the 2024 

 the UN’s commitment to “enhance international governance of AI for the benefit of humanity.”

, released just ahead of the Global Dialogue on AI in Geneva earlier this month, reveals a tension at the heart of the UN’s wider global AI governance work. While the report recognizes the growing concentration of AI power in the hands of a few private corporations and countries, it simultaneously advocates for the rapid global diffusion of AI to unlock benefits in sectors such as health, education, food security while also promoting economic productivity, 

. Achieving this vision depends on significant investments and contributions from the private sector to develop AI infrastructure globally, to support nations’ compute capacities, provide technical expertise, and cultivate enabling regulatory frameworks. In doing so, the UN risks reinforcing the very market structures and infrastructural dependencies on Global North Big Tech corporations that have led to current concentrations of power.

UN Secretary-General António Guterres introduced the IISPAI, which operates under a "

strictly scientific, non-political mandate

first global, fully independent scientific body dedicated to helping close the AI knowledge gap

.” Yet the issues it addresses are not purely scientific; the questions it tackles in its pursuit of closing AI knowledge gaps, including those about how states should build AI capacity/compute, attract private investment, develop talent, and govern AI systems are fundamentally political. The IISPAI’s commitment to “

balanced analysis of the risks and opportunities of AI

”—defined as evaluating empirical evidence “

without undue bias towards optimism or pessimism

”—does not remove these political dimensions. Decisions about what constitutes an AI opportunity, what counts as a risk, and which governance pathways should be prioritized inevitably reflect underlying political assumptions about the role that AI systems should play in global society. Despite the IISPAI’s claims to independence, 

, the Panel’s assumptions about the benefits and risks of global AI adoption are consistent with a broader trajectory in UN AI governance over the past several years.

, which traces the UN’s AI governance discourse since 2019, the UN has consistently advanced an economic development approach to AI governance that frames AI adoption as a prerequisite for development and digital inclusion, while privileging private sector investment, cross-border data flows, and AI capacity building and diffusion in Global South emerging markets. These arrangements threaten to reinforce global reliance on Global North digital infrastructure. The primary concern driving the UN’s AI governance efforts has not been whether AI 

 be adopted. The UN treats the global spread of AI as inevitable and necessary. Instead, the UN is focused on determining how more countries can participate in the global AI economy while 

If data was the fuel of the digital economy, compute—the computational infrastructure needed to develop and deploy advanced AI systems—is the 

. While the report points to evidence about AI systems and the inputs they rely on—such as computing power, data, and engineering talent—being concentrated in “

a handful of firms in a handful of countries

,” the report simultaneously promotes local infrastructure development which involves 

. However, private investment, even when undertaken by domestic corporations, 

 into AI sovereignty. A country may build domestic data centers through domestic firms that still rely on NVIDIA chips, 

, and OpenAI models. Such infrastructural entanglements increase local AI capacity without necessarily increasing state’s domestic control over AI.

 that “Countries that rely on foreign models, cloud infrastructure and data pipelines may gain access to AI while losing practical control over its standards, safeguards and local fit.” It also acknowledges that developing advanced AI capabilities is increasingly beyond the reach of individual states, requiring “coalitions of countries or major stakeholders” to pool resources such as data, capital, compute, energy, and talent. Yet the Panel's proposed response—a UN-facilitated 

—does not fully resolve this reliance on foreign infrastructure. To the extent that the fund depends on financial or 

 from dominant technology firms, including compute access, technical expertise, AI models, and training resources, it risks reinforcing the very dependencies it intends to overcome. The Global AI Fund may grant countries greater access to AI while remaining reliant on the tech corporations that control digital infrastructures on which those capabilities depend.

The UN’s global AI governance discourse appears to simultaneously throttle the gas and brakes on AI, promoting global AI adoption while it warns about the serious risks it poses to society. The UN must, at some point, acknowledge that the global diffusion of AI technologies and challenging AI market concentration are fundamentally different governance objectives. Rapid global AI adoption necessarily involves more global users, more investments in private infrastructure, and more global participation in an AI economy currently dominated by Big Tech corporations.

The IISPAI does deserve credit for confronting the political economy of AI more directly than previous UN AI governance documents. Yet even as it identifies dependency and concentration as central challenges, it treats nations' exclusion from the global AI economy as the problem. The IISPAI can identify private concentrations of compute, data, talent, and frontier AI development as challenges for global governance, but it fails to recommend reforms that would directly redistribute control over those resources.

In this respect, the IISPAI report reflects a broader pattern within UN AI governance which has become increasingly willing to critique the political economy of AI development and the global power relations it produces, yet remains reluctant to advocate for structural reforms that would challenge existing distributions of technological and economic power. Earlier reports, including 2023 “

,” similarly warned about the concentration of AI capabilities among a small number of private corporations. Yet these concerns were largely absent from the adopted text of the 

. Of course, the UN is institutionally incapable of enacting binding resolutions that might redistribute control over the development of AI systems. But it could, at minimum, confront the contradiction of condemning AI market concentrations while advancing a model of global AI development that reinforces the very corporations and digital infrastructures at the center of those concentrations.

Questions of global AI governance are not merely technical questions answered by ‘politically neutral’ empirical evidence; they are 

questions about whose interests are served by particular governance arrangements. Decisions about 

governance pathways are considered feasible inevitably reflect assumptions about AI development, innovation, and power. The question facing the UN is therefore not simply how to ensure that more countries can benefit from AI adoption, but whether it is willing to imagine forms of governance that redistribute control over the infrastructures on which AI depends. As AI governance shifts from debates over ethics to debates over infrastructure, international organizations like the UN, will increasingly have to decide whether their role is simply to expand participation in existing AI markets or to reshape the distribution of power that underpins them.

The harder question, therefore, is not whether the UN can diagnose concentration, but whether it is prepared to act upon what its own analysis reveals. Until then, the UN’s AI agenda may remain caught in a cycle of critiquing concentrated power over AI systems while simultaneously reproducing the conditions that sustain it.

European Union flag displayed on a screen and AliExpress, Temu and Shein icons displayed on a phone screen are seen in this illustration photo taken in Krakow, Poland on July 1, 2026. (Photo by Jakub Porzycki/NurPhoto via AP)


 for failing to adequately assess and mitigate the risks posed by illegal, unsafe and counterfeit products on its platform — the largest penalty issued to date under the Digital Services Act, and the third overall.

AliExpress, which had 193 million European users last year, called the decision disproportionate and said it is reviewing its options.

“AliExpress has been and continues to be committed to meeting our obligations to consumers. We invest substantial resources in risk assessment and mitigation, product safety and consumer protection,” AliExpress said in a statement to Tech Policy Press. The company said it had worked constructively with the European Commission in the spirit of cooperation and transparency, but the fine ignored its “sound risk management framework and the significant, proactive enhancements.” 

The Commission said AliExpress failed to properly evaluate whether it had enough resources to monitor potentially illegal products, overstated the effectiveness of its detection and removal systems, and did not adequately consider how its recommendation and advertising tools could increase the visibility of illegal goods.

Investigators found examples of counterfeit products, unsafe toys and potentially dangerous cosmetics remaining available on the platform after being flagged.

“The spread of counterfeit clothing, unsafe toys, dangerous cosmetics and other illegal and harmful products is not an unavoidable cost of shopping online,” said Henna Virkkunen, the European Commission’s executive vice president for tech sovereignty, security and democracy.

“Scale is not an excuse; risks must be identified and addressed systematically to ensure consumers can safely shop online,” she added.

The AliExpress decision reflects a broader shift in the EU’s enforcement focus under the DSA, moving beyond social media platforms toward online marketplaces that facilitate the sale of physical goods.

Early DSA cases largely centered on content and transparency issues on platforms such as X and Meta. The AliExpress and Temu cases involve products that regulators say may pose direct safety concerns for consumers.

The European Board for Digital Services identified illegal products sold through online marketplaces as one of the most persistent systemic risks across designated platforms in its 

The report highlighted challenges including sellers who attempt to evade platform controls by modifying listings or returning under new identities after products are removed — a problem also cited by the Commission in its AliExpress investigation.

The regulatory challenge for marketplaces is therefore not only to remove individual listings but also to demonstrate that their systems can identify and prevent recurring risks at scale.

All three major Chinese e-commerce platforms operating in Europe are now facing some form of DSA oversight.

, must submit an action plan to the Commission by Aug. 28. Shein is currently facing 

 over alleged shortcomings involving illegal products, addictive design features and recommendation system transparency, although no fine has been issued. One in five Europeans shops from the three platforms at least once a month, Virkkunen told reporters.

The enforcement actions are taking place alongside separate EU trade measures affecting low-value imports. A 

 on parcels valued below €150 from China took effect July 1, targeting a different part of the same cross-border commerce system.

The AliExpress fine does not conclude the DSA process.

 outlining how it will address the Commission's findings. The Board will then provide an opinion, followed by a Commission decision setting out the implementation timeline. If AliExpress fails to comply with the final requirements, the Commission can impose additional periodic penalty payments.

The process will test how the EU's enforcement model works after penalties are issued. In the case of X, the company 

 following its December fine covering advertising transparency and researcher data access. The Board raised concerns that some proposed measures were insufficient, but the Commission 

 after additional clarifications. The case highlighted the separate roles of the Board, which provides independent opinions, and the Commission, which makes final enforcement decisions — and the gap that can open between the two.

AliExpress also has prior form here. It's 

 with the Commission have resolved several open investigation strands covering recommender systems, advertising transparency and trader traceability. Those commitments did not prevent Monday's fine, which covered broader concerns about the platform's risk assessment and mitigation systems. The Commission found the commitments lacked specific, measurable targets and external verification mechanisms.

The Commission also noted that the relatively early stage of DSA enforcement was taken into account when calculating the penalty — meaning future cases could involve higher sanctions.

The next test is whether AliExpress’s corrective measures deliver changes beyond written commitments. The case highlights a central feature of the DSA, which focuses not only on whether illegal content or products are removed but also on whether platforms have systems capable of identifying and reducing risks before harm occurs.

The approach places greater emphasis on internal governance, risk assessments and transparency around platform operations. The Commission has also pointed to the role of external scrutiny in identifying risks. Consumer groups and researchers have raised concerns about products sold through online marketplaces, including 

, the European consumer organization, that contributed to the Commission’s proceedings against Temu.

The Board’s second annual report found that outside actors consistently identified risks that platform self-assessments missed, a gap that AliExpress’s October plan will need to address.

AliExpress must submit its remedial action plan by Oct. 20. The Board will have one month to issue its opinion, followed by a further month for the Commission to adopt a final decision and establish an implementation timeline.

Temu's earlier deadline on Aug. 28 will provide the first indication of how the Commission evaluates post-fine compliance plans from marketplace platforms — and how much weight the Board's opinion carries when the two institutions disagree.

Shein remains under investigation. Any further action would add another major compliance process to the Commission's workload as it continues building the enforcement framework for the DSA. Whether that process produces structural change or paper commitments is what the second half of 2026 will show.

This article was updated to include a comment from AliExpress.

A group of shareholders of Samsung Electronics Co. hold a rally criticizing a tentative wage agreement struck between the company and its labor union near the home of Samsung Electronics Chairman Lee Jae-yong in Seoul, South Korea, Thursday, May 21, 2026. The letters read "Unjustified transfer strike." (AP Photo/Ahn Young-joon)

Workers in Asia Are Fighting for Protections as AI Threatens Jobs

A police officer checks the list of a polling station during the parliamentary elections in capital Nicosia, Cyprus, Sunday, May 24, 2026. (AP Photo/Petros Karadjias)

EU Political Ad Rules Leave Elections Open to Foreign Manipulation

June 22, 2026—Andy Burnham with colleagues from the Parliamentary Labour Party in Westminster Hall at the Houses of Parliament in central London, as he returns to the House of Commons to take up his seat after winning the Makerfield by-election. (Yui Mok/Press Association via AP Images)

How Will Andy Burnham Handle Tech Policy as UK Prime Minister?

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