Regarder un film, voir une exposition, assister à un concert, débattre, boire un café, lire, acheter ou déposer des livres... Voici ce que propos l'Akvarius, unique oasis dans le désert culturel de Mitrovica-Nord. Reportage.
- Articles / Kosovo, Tensions Nord Kosovo, Vreme (Serbie), Culture et éducation
Life of factory workers in China... pic.twitter.com/OhZIuHipUQ
— Mayte Chummia (@Maytechummia) January 2, 2026 Geez. I can't even imagine working in a place like that. Maybe there are spots in the West that mirror that shit but I kinda get all those suppressed reports of Chinese workers going postal (showing my age with that phrase huh).Sounds real close to full capitulation.
Are the hawks so hungry to flex muscle that negotiation is no longer possible? I hope not because if we're able to get what we want then its as close to perfect victory as anything I've seen in the past two decades.
AI risk pyramid. Source: European Commission https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai
Bao-Chau PhamWhen the European Union‘s Artificial Intelligence Act finally entered into force on 1 August 2024, it was widely described as a landmark: the world’s first comprehensive framework for regulating AI. Throughout the policy-making process from early 2018 onwards, much of the debates focused on how the Act regulates – its risk categories and list of banned use-cases – and what these legislative choices do. This included commentary on its neglect of human rights, as well as concerns that it might stifle innovation. Sitting underneath these discussions, however, is a simpler question: what problem is the AI Act actually trying to solve?
That question is at the heart of my article, co-authored with Sarah R. Davies and published in Critical Policy Studies. As the EU now debates delays and possible revisions of the Act – or “rollbacks”, as more critical voices have put it – the question feels newly relevant. The recently proposed Digital Omnibus package may not just tweak the AI Act as a legal instrument. We may be witnessing the re-articulation of the very problem the legislation was designed to address.
In this post, I briefly introduce the argument of our article, reflect on why this way of reading the AI Act matters, and suggest why current discussions about revisiting the Act make the question ”What problems is the AI Act really solving?” even more salient.
Reading policy as problem representations: the WPR approach
Our article starts from the premise that policies do not merely respond to external issues. Instead, policies actively participate in producing the very problems they are designed to solve. Drawing on Carol Bacchi’s seven-step “What’s the Problem Represented to Be?” (WPR) approach, we therefore read the AI Act as a document that constructs a particular understanding of AI and its governance.
From this perspective, asking whether the AI Act is effective or proportionate is only half the story. The more fundamental question is how the Act frames AI as a policy problem in the first place. This, in turn, allows us also to interrogate what and who gets foregrounded, sidelined, or taken for granted in the process.
Using the WPR approach, we identify two dominant, yet ambivalent, problem representations of AI in the Act. On the one hand, AI is framed as a major economic and societal opportunity, essential for competitiveness and Europe’s (digital) future and something that Europe risks „missing out“ on. On the other hand, AI appears as a source of significant risk, particularly to European fundamental rights and democratic values.
This dual framing is not a contradiction to be resolved; it is the organising logic of the Act. The cascading scale of risk – dividing systems into minimal, limited, high, and unacceptable risk categories – emerges directly from this way of problematising AI. The AI Act frames the policy problem as one of managing trade-offs: how to promote uptake while preserving trust, or how to govern a fast-moving technology without stifling it.
Seen this way, the AI Act is not just regulating technology. It stabilises a particular vision of AI as something that can, and should, be rendered governable through categorisation, technical requirements, and legal obligations. Other possible ways of characterising AI as a policy issue, as well as other possible responses, are, in the process, foreclosed.
Importantly, the problem is also represented in a way that positions the European Union as the primary actor capable of solving it. One of our key arguments that follows is that the AI Act also plays a role in enacting a particular version of Europe and Europeanness. It constructs and institutionalises the notion of the EU as an exceptional, morally authoritative policy actor in global AI governance. In this sense, the AI Act is as much about Europe’s self-identification in a global technological landscape as it is about regulating specific AI systems.
Why this matters now: shifting goalposts
At first glance, current discussions about revising the AI Act may look like implementation politics: delayed technical standards, pressures from industry and lobbyists, and geopolitical concerns about competition. If we return to the question of problem representation, however, these debates take on a different meaning. They point to a possible shift in what is understood to be the central problem that AI policy should address.
As we argue in our article, during the AI Act’s legislative process the dominant discourse centred on risk to fundamental rights alongside economic opportunity. Increasingly, however, public debate is framed in different terms: how to avoid over-regulation, maintain competitiveness, and keep pace with global AI development. The risk that now receives the most attention is not always harm to citizens or democratic institutions, but harm to innovation ecosystems and market position.
This does not mean that fundamental rights have completely disappeared from the conversation, but it does suggest that they may no longer be the primary lens through which the policy problem is articulated. From a WPR perspective, the question is not whether the AI Act is being weakened or strengthened. It is whether the problem the policy is meant to solve is being re-articulated.
If AI is increasingly represented as a competitiveness challenge rather than a rights challenge, then different policy solutions, which favour scalability and speed, follow suit. Seen in this light, current proposals to simplify the AI Act are interventions in an ongoing struggle over how AI should be understood as a matter of public concern and whose interests should take priority when trade-offs are made.
This is precisely why the question “what problem is the AI Act solving?” remains important. It reminds us that regulation and policy are never only about technicalities. If we take problem representation seriously, then revisiting the AI Act is equally about deciding which understandings of AI become stabilised in European governance going forward.
To conclude, the AI Act was never just a response to extraneous technological developments. From a critical policy perspective, it reads as an attempt to stabilise a particular way of thinking about AI and governance in Europe. As the constellation we saw in the AI Act is now being reconsidered, returning to the question of problem representation can help us unpack and trouble what is at stake. Whether the AI Act ultimately changes or not, it is worth remembering that the debates about AI governance are not only about how we regulate, but about what we think needs regulating, why, and for whom.
Bao-Chau Pham is a recent PhD graduate in Science and Technology Studies from the University of Vienna (Austria). In her dissertation, Bao-Chau explored imaginaries of artificial intelligence in European policy and media discourses.
The post Revisiting what problems the EU AI Act is actually solving appeared first on Ideas on Europe.
Az Air Base blogon, ha nem is túl gyakran, de időről időre előveszem a hajózás témáját is, (ami olvasóim körében, számomra is némiképp meglepő módon, kedvező fogadtatásra talált). A kereskedelmi hajózásra leszűkítve a kört volt már szó az SS Baron Gautsch katasztrófájáról, a Greenwichben kiállított Cutty Sark klipperről, a XVIII. századi holland Amsterdam vízen úszó, jól sikerült replikájáról, spliti, fiumei és londoni múzeumok gyűjteményéről vagy csak egyszerűen egy-egy tengeri kikötő forgalmáról. Így lesz ez most is, a tavaly nyáron készült fotóimból összeállított album formájában.
Debut of the Taiwan F-16V Block 70! pic.twitter.com/QJAJlvEEfl
— Owen | Aviation Photographer (@OwenAviation) December 21, 2025