Like any other commercial sector, the fashion industry is also working towards reducing the risk of ‘unsuccess’, that is decreased sales and loss of brand attractiveness and value. One way to achieve all this is by mining the large quantities of text and data that are shared daily through social media, in order to ‘understand’ consumers’ tastes and desires well before they are externalized through purchasing decisions. In other words, the goal is to anticipate such decisions. As a result, ‘algorithm fashion’ is on the rise. From a technical standpoint, however, mining processes often – though not invariably – require the undertaking of acts of reproduction of content – whether image, video, or text – that might be inter alia protected by copyright and/or related rights. As such, these processes become relevant from a copyright perspective too. This chapter considers the rise of algorithm fashion from an EU perspective, focusing in particular on the scope of the recently adopted exception or limitation for text and data mining found in Article 4 of the DSM Directive (2019/790) in order to determine whether and to what extent unlicensed mining activities may be undertaken in a commercial setting.

A key principle of EU law is that of direct effect. The now CJEU recognized it for the first time in the landmark 1963 judgment in Van Gend en Loos (Case 26–62, EU:C:1963:1).Insofar as EU directives are concerned, the dominant view is that they do not produce any horizontal direct effect. This means that they cannot be relied upon in private-party proceedings and do not impose obligations on individuals. Nevertheless, EU directives may produce—at certain conditions—a vertical direct effect. In cases in which EU law imposes upon Member States ‘the obligation to pursue a particular course of conduct, the useful effect of such an act would be weakened if individuals were prevented from relying on it before their national courts and if the latter were prevented from taking it into consideration as an element of [EU] law’ (van Duyn, Case 41–74, EU:C:1974:133, para 12).

The training of Artificial Intelligence (AI) models relies on extensive amounts of “data,” often sourced from content protected by copyright, related and sui generis rights. The discussion of whether and how to strike a balance between licensing and exceptions under copyright law is one of global relevance. While some countries have adopted or considered adopting specific exceptions to allow text and data mining (TDM), others (most) have not introduced any new legislation. In Europe, much of the attention has so far centred on Article 4 of Directive 2019/790 (DSMD), including in the context of a potential UK reform.The starting point of this contribution is the following four-fold observation. First, TDM may be part of AI training processes, but it is neither synonymous with AI training nor is it all that AI training entails, including in terms of acts restricted by copyright and related rights. Second, from a European (thus including both the EU and the UK) perspective, limiting the attention to Article 4 DSMD is myopic, as national case law demonstrates. Third, calls have recently been made to relax EU copyright rules to facilitate “research,” seemingly including the President of the European Commission herself, who announced forthcoming legislative proposals “to make Europe the home of innovation again.” Fourth, the UK Government’s Copyright and AI consultation has recently ended: should no reform be ultimately undertaken, the application of the existing TDM exception will depend to a large extent on how courts construe the notions of “research” and the “non-commercial” requirement thereof.Moving from the above, this study investigates whether and to what extent unlicensed AI training activities could be undertaken by relying, not on Article 4 DSMD as transposed into national law or a hypothetical reform of the UK system of exceptions, but rather on what appear to be so far potentially overlooked defences. Reference is made specifically to research and education exceptions, notably Article 3 DSMD and Article 5(3)(a) of Directive 2001/29 (InfoSoc Directive), also read in light of Article 5 DSMD. The discussion of other jurisdictions – including the US and countries, like South Korea and Singapore, which have adopted open-ended fair use-style defences – is also undertaken. This is done to determine whether unlicensed AI training, including training seemingly done for the purpose of research or education/learning, might be considered lawful.In light of the context summarized above, the study tackles two key questions: (a) whether unlicensed AI training may be classified as “research” or even “learning” in the context of “teaching,” and (b) whether commercial AI developers may take advantage of the provisions above. Ultimately, both questions are answered in the negative, finding that no exception or open-ended defence fully covers unlicensed AI training activities. As a result, a licensing approach (and culture) appears to be the way for AI training to be undertaken lawfully, including when this is done for “research” and “learning.”

Any attempt to define the phrase ‘cultural heritage’ is riddled with difficulties and ambiguities, considering that the very notion of ‘culture’ proves challenging to grasp even in its basic contours. That said, broadly speaking, cultural heritage refers to things – whether tangible or intangible – that may be movable or immobile (as well as underwater) and have artistic, historical, archaeological, ethnological/anthropological, scientific, and social value. It is expected that for things to ascend the status of cultural heritage, the passing of a rather significant period of time post-creation is a given, with the result that ‘natural’ forms of legal protection vested in them, such as copyright, might have lapsed in the meantime. But is cultural heritage ever in the public domain? Any answer in the affirmative would overlook both substantive rights – including under intellectual property law (think of trademarks or geographical indications) – that are not subject to the same expiration date as copyright, and other forms of legal protection that might have the practical effect of preventing any actual fall into the public domain.The expectation that cultural heritage might be beyond anyone's control also proves incorrect if the reference is made to a country rich in cultural heritage such as Italy. The Italian Cultural Heritage Code (ICHC) does indeed set forth the general principle according to which for-profit uses of cultural heritage under the responsibility of the Italian state are subject to a requirement of preventive authorization of the competent public administration. Articles 107 and 108 ICHC provide that a public administration, for example, the state museum where the object in question is being held – be it Michelangelo’s David (on display in the Galleria dell’Accademia in Florence), Sandro Botticelli’s Birth of Venus (also in Florence, though in the Uffizi collection), or Leonardo’s Vitruvian Man (found in the Gallerie dell’Accademia in Venice) - can authorize third-party uses of the image of an object belonging to Italian cultural heritage, subject to the payment of royalties. In essence, the ICHC restricts the unauthorized, for-profit use of the image of a cultural heritage artefact. First adopted in 2004, the ICHC has remained largely dormant (and thus overlooked) for a long time. Things have changed over the past few years, that is (a) since some Italian state museums have started taking legal action to repress unauthorized uses of the image of cultural heritage artefacts they hold and (b) courts have issued decisions that are not only in favor of the museums but also establish potentially far-reaching legal principles. Two in particular stand out: the first is that cultural heritage artefacts enjoy image rights – to use US legal language: publicity right – protection in the same way as natural persons do; the second is that the ICHC would not only apply to uses of Italian cultural heritage on the Italian territory, but would extend to uses taking place anywhere, well beyond the Italian borders, including potentially in the US.This article details the relevant jurisprudential evolution before providing some critical remarks regarding the territorial scope of application of the ICHC and the compatibility of the Italian law with European Union (EU) law. An attempt to map the way forward, including for US companies that are using or are interested in using images of Italian cultural heritage, is also made.

The analysis of liability aspects facing Artificial Intelligence (‘AI’)-generated outputs under copyright and related rights has been overlooked compared to other issues connected to the development and use of AI. This study fills this gap by exploring pertinent questions under international, EU and UK law. Specifically, the study tackles actionable reproduction, allocation of liability, and availability of defences. The analysis ultimately shows that, while it is clear that each case will need to be decided on its own merits, the generative AI output phase raises several profiles of liability under copyright law. If the goal of policymakers and relevant stakeholders is to ensure the balanced and sustainable development of AI, then the issues related to the generation and dissemination of AI outputs need to be given ample attention and a greater role in the debate than what has been the case so far, whether it is in the context of risk assessment and compliance, licensing initiatives, or in contentious scenarios.