The internet, once celebrated as a decentralized public sphere, is increasingly undermined by practices such as generative search and shadow banning, which divert traffic and suppress visibility. Generative search, powered by Retrieval-Augmented Generation (RAG), synthesizes content into direct answers, bypassing websites and depriving them of traffic and revenue. This threatens the sustainability of independent content creators, small enterprises, and the open web ecosystem. Shadow banning, a practice that intentionally reduces the visibility of social media posts through algorithmic moderation, exacerbates these issues by chilling free expression and limiting transparency and accountability.

This article explores these opaque practices through a legal and regulatory lens. The first part examines the rise of generative search, analyzing its technological and legal implications, including copyright infringement, unfair competition, and unjust enrichment. It also evaluates potential solutions such as licensing agreements and agentic AI. The second part focuses on shadow banning: algorithmic dissuasion, de-ranking, and the reduction of traffic, with specific attention to China’s Regulation on Algorithmic Recommendations (RAR) and the EU’s Artificial Intelligence Act (AIA). Both frameworks offer partial solutions but fall short of ensuring fairness, transparency, and redress mechanisms.

Ultimately, the shift toward centralized control by dominant platforms prioritizes profit and risk management over innovation, fairness, and diversity in online expression. To counteract these trends, regulatory interventions, algorithmic transparency, and equitable frameworks are essential. Without such measures, the internet risks losing its character as a democratized public sphere for free expression and innovation.

The use of copyrighted works in training large AI models has sparked numerous lawsuits globally. This Article examines China’s evolving regulatory landscape, and analyzes two academic proposals for China’s Artificial Intelligence Law, identifying key areas of divergence and consensus regarding the fair use of copyrighted works in AI training. By comparing three different legal approaches to characterizing AI model training, this Article argues that this process qualifies as fair use. This is because machine learning leverages vast corpora to internalize underlying linguistic and creative patterns, rather than storing or directly reproducing the protected works. As a result, the use of copyrighted material in the training phase qualifies as incidental reproduction and transformative use, which, according to our empirical study, does not unreasonably harm the legitimate rights and interests of copyright holders. Furthermore, given the market failure in AI model training licensing, this Article contends that recognizing AI model training as fair use better aligns with China’s legal framework and the practical needs of technological development. To ensure legal certainty, this Article proposes introducing a machine learning exception within either the ongoing revision of the Regulations for the Implementation of the Copyright Law, or future AI legislation in China.

Choice of law clauses are a central instrument for safeguarding party autonomy. At the same time, it must be ensured that consumers are protected and sufficiently informed, particularly with regard to the applicable law. The CJEU has established important principles in this regard in the Amazon decision. Based on the principles of choice of law in private international law, the article takes a closer look at the protection of the weaker party. The Amazon decision is placed in the system of private international law and justified by taking into account the various functions of information duties.

In this Article I suggest that, due to the changing nature of the polarization of Western societies, another important factor should be taken into account in assessing the relationship between public trust in the judiciary and judicial independence, namely court communication with various segments of the society, including the ordinary people. More specifically, my key argument is two-fold. First, we live in the disinformation age and apex courts can easily be portrayed as detached from the ordinary people. This endangers their social embeddedness, which in turn might increase the likelihood of the acceptance of court-curbing. Second, apex courts should proactively respond to this challenge by adopting embedding strategies aimed at all segments of society, and not just the elites. I identify four such strategies: (1) The media strategy; (2) proactive engagement with the precariat via “reaching out” activities such as social events and holding hearings outside the courts’ seats; (3) minimalization of controversial off-the-bench activities of judges; and (4) self-awareness and avoidance of structural judicial bias. Finally, I discuss the risks and limits which courts and judges face when they engage in these embedding strategies.

The emergence of generative artificial intelligence (generative AI) has profoundly reshaped content creation by lowering marginal costs, altering the role of human creators, and restructuring industrial processes. These shifts pose three critical challenges for intellectual property (IP) frameworks: redefining creative behaviors in human–AI collaborations, sustaining incentives for innovation, and addressing gaps in copyright mechanisms. Against this backdrop, the Beijing Internet Court (BIC) issued China’s first judicial ruling on the copyrightability of AI-generated images in 2023. The court held that AI-generated content (AIGC) could qualify for copyright if there is demonstrable human intellectual input and originality. However, the ruling also emphasized the need for case-by-case assessments, particularly in hybrid human–AI creative processes. This landmark decision has sparked intense debate over whether and how AIGC should be considered as a “work” under copyright law, with particular focus on requirement of human authorship and the threshold of originality. In addition, a systematic analysis of the AIGC value chain highlights the far-reaching implications of human authorship claims for various stakeholders, including creators, AI developers, prompt engineers, and end users. To address these emerging legal and ethical issues, this Article proposes two key reforms: first, extending the scope of fair use under Article 24 of China’s Copyright Law to account for AI-generated works, and second establishing a “whitelist” mechanism in the Regulations for the Implementation of the Copyright Law. These measures aim to balance the protection of human creativity, ensure equitable value distribution among stakeholders, and foster sustainable innovation in the AI-driven creative ecosystem.

A major challenge for contemporary legal constitutionalism is a crisis of public ethics that manifests in the lack of mutual toleration and institutional forbearance towards the judiciary. To showcase the importance of these norms in the relationship among co-equal branches of government, I focus on three cases, one where these norms have been present—South Africa—one where they have been absent—Mexico—and one case in between—United States. Until this crisis is addressed, the authority of apex courts will continue to be under threat. The Article suggests that a starting point to address the public ethics deficit may lie in shifting comparative constitutional law scholarly attention to the political sphere.

Although choice-of-jurisdiction clauses are routinely enforced by courts in the United States, there are circumstances where they are subject to special scrutiny. One of these circumstances is when the party resisting the clause was not provided with proper notice as to the existence of the clause or the identity of the chosen jurisdiction. This Article first reviews the existing case law in this area and shows that while some U.S. courts have refused to enforce clauses for lack of notice, others do so as a matter of course. It then discusses several decisions where U.S. courts have held that notice may serve as a substitute for consent to bind parties to choice-of-jurisdiction clauses in agreements that they never signed.

Democratic decision-making processes tend to take less account of future interests than of present ones, thereby jeopardizing not least the foundations of future democracy. Democratic short-termism causes considerable problems, especially when decisions made now have serious consequences for the future and can hardly be reversed. Thus, the current climate protection legislation, which is far too weak worldwide, threatens to impose unbearable and unjust burdens on people in the future and deprive them of any political leeway for shaping their own policies, thus also undermining the very basis of democracy. In general, it is one of the functions of constitutional law to help overcome such short-termism of decision-making processes. However, in the case of climate protection, this is difficult because the veto power of constitutional law and constitutional courts has comparatively little impact here—for reasons related to the factual characteristics of climate protection. Nevertheless, constitutional law and courts have their own potential in climate protection that needs to be further developed in order to overcome some of the democratic short-termism.

The rapid advancement of generative artificial intelligence (GenAI) presents new challenges and implications for anti-discrimination law. While much of the legal discourse so far has focused on general AI governance, there is a growing need to recognize that different technical models—ranging from algorithmic decision-making systems (ADS) to GenAI—pose distinct risks and require tailored regulatory responses. This Article serves as a foundational introduction to this issue, highlighting the diverse ways in which GenAI can both reinforce and mitigate discrimination and where (European) anti-discrimination law stands in the face of these challenges. The primary goal is to raise awareness of the crucial interplay between AI architectures and legal frameworks. By sketching the underlying technical and structural mechanisms this Article aims to foster an understanding of the potential harmful impact GenAI can have, and how this is met by an anti-discrimination law, not yet equipped for these new “actors.” Thus, it will reveal weaknesses that will need to be addressed by research and policy alike.

The enforcement of forum selection and arbitration agreements against consumers and other parties in disadvantaged bargaining positions has significant consequences for access to justice. As a result, some legal systems simply decline to enforce jurisdictional agreements against certain groups. This is not the case in the United States. To the contrary, such agreements enjoy a strong presumption of enforceability across the board. As a result, the ability of individual parties to secure remedies for violations of their legal rights is significantly curtailed. Because private enforcement plays such an important role in state and federal regulatory regimes, impairing the ability of individual litigants to sue also erodes the accountability of corporations for violations of law. This Article begins with the understanding that jurisdictional agreements constitute contractual waivers of rights. This highlights the two primary sources of law that can be used to police them: contract law and “anti-waiver” rules and doctrine. The Article considers each in turn. It concludes that contract law no longer provides a meaningful constraint on the use of adhesive jurisdictional agreements, but that anti-waiver rules at the state level may.

In recent years, courts across the world have increasingly held governments accountable for addressing climate change. While such rulings have fueled optimism about constitutional law as a vehicle for climate ambition, this Article argues that the role of constitutional law in advancing climate goals is far more complex and contested. Constitutions encapsulate diverse and sometimes conflicting values, which can create tensions when courts adjudicate climate policies. As government climate measures become more concrete, conflicts arise between rights, institutional structures, and political realities. Drawing on examples from Germany, Canada, and Mexico, this Article highlights the challenges of adjudicative uncertainty, the underspecificity of constitutional norms, and the polyvocality of constitutional values in the context of climate change. This Article concludes with recommendations for judges to adopt a principled, context-sensitive approach to constitutional climate adjudication, balancing the urgency of climate action with the complexities of state capacity and constitutional structures.

For EU jurisdictions, Regulation (EU) No 1215/2012 on jurisdiction and recognition and enforcement of judgments (the Brussels I Regulation) and the 2005 Hague Choice of Court Convention (the 2005 Choice of Court Convention) are the pivotal instruments to assess “consent to jurisdiction” in a choice of court agreement. This Article examines the symmetries in both instruments. The relevant rules combine inferred consent with a residual reference to the law of the chosen court, as opposed to providing a uniform substantive standard on the notion of “consent” to jurisdiction. The practical operation of this hybrid solution is examined by reference to the relevant case law of CJEU and selected domestic courts. The Article then considers specific matters where consistency between the two regimes is uncertain. In conclusion, while alignment on consent to jurisdiction under the Brussels I Regulation and the 2005 Choice of Court Convention is desirable, it is expected that autonomous standards on consent to jurisdiction will be consolidated within the respective scopes of application of the instruments under review.

This Article discusses China’s content moderation in the age of artificial intelligence. It first introduces two long-overlooked features of China’s content moderation: the medium-based model and the “No-Dispute” Policy. The former emphasizes that content moderation in China varies based on different media, while the latter argues that China’s content moderation is often content-neutral rather than being driven by ideology or having an official stance. The Article then summarizes the three main challenges artificial intelligence presents to content moderation: a shift in structure from the traditional “state v. citizen” dichotomy to the “platform–government–citizen” triangle; a transition in means from human review to algorithm-based and machine-based moderation; and stimulating a reimagination of traditional theories and doctrines of freedom of speech in terms of standards and classification. Finally, the Article takes online violence, one of the most prominent issues in contemporary Chinese content moderation, as a case study to examine specific issues in China’s content moderation in the era of artificial intelligence.