Periodic Reporting for period 1 - HUMANads (Human Ads: Towards Fair Advertising in Content Monetization on Social Media)
Période du rapport: 2022-07-01 au 2024-12-31
Human ads are Internet influencers who earn revenue by creating and monetizing authentic and relatable advertising content for their armies of followers, by relying on business models such as influencer and affiliate marketing. This content often results not only in commercial but also political (hidden) ads, which look the same, are posted by the same persons, are displayed in the same digital space, to the same audiences, and raise the same transparency issues. In this environment, consumers and citizens can no longer distinguish between ads and non-ads, and between commercial and political communications. They are faced with a double transparency problem: (i) human ads have incentives to hide commercial interests, and (ii) platforms have incentives to algorithmically amplify human ads engagement in opaque ways. This reflects a general good faith and fair dealing problem: the social media economy is increasingly based on deceit, which leads to new forms of vulnerability for consumers and citizens on digital markets. Given its complexity and relative novelty, this phenomenon has not yet been the object of sustained academic or regulatory inquiry. HUMANads tackles this comprehensive research gap by exploring why a general European legal regime on fair advertising by human ads on social media platforms is necessary, and what it would entail. First, it articulates new theory of fair advertising in EU consumer law, in the context of content monetization by human ads across commercial and political speech. Second, it gathers evidence relating to business models, advertising prevalence and legal uncertainty through innovative interdisciplinary methods including digital ethnography, comparative law and natural language processing (NLP). Third, it proposes criteria for the assessment of resulting consumer harms, and translates them into a new normative governance model mandating more stringent transparency obligations on social media platforms.
The HUMANads project has been off to a great start, and has already achieved many of its objectives in the first two years. In a nutshell, with a team of 5 researchers, we completed 19 publications (10 peer-reviewed articles; 1 edited volume; 5 non-peer-reviewed articles; and 3 policy briefs), we organized 15 events (workshops, conferences and outreach events), and participated in 49 events as invited participants or selected speakers. All these activities have been taking place across the envisaged three fields essential in the project: law, computer science and media studies. The project has been enjoying an ever-growing network ranging across Europe, North America, South America and Australia, and is currently seen as a leading research group on the topic of content monetisation, influencer marketing and platform governance. This research has also been heavily featured in mass media (e.g. TIME, Business Insider, Financial Times), and its computational methodologies are already applied in practice by advertising self-organisations such as the Stichting Reclame Code (NL) in its Influencer Monitor. Moreover, the legal findings of the project have been used in a collaboration with the European Commission (DG Justice and Consumers) that led to the creation of the Influencer Legal Hub, a first of its kind multimedia resource on influencer marketing and European consumer protection.
The project has generated innovative input in legal and computational methodologies. In law, the project has thoroughly explored the application of consumer protection legislation and platform liability implications to the influencer marketing context. Different publications (e.g. the forthcoming publication in the Internet Policy Review) also entail multidisciplinary methodologies such as the walkthrough method, derived from media studies and applied to legal studies. Lastly, in computer science we were able to propose a novel methodology for the discovery of influencers and the detection of hidden advertising, and to even label a dataset based on public registrations to the Dutch Media Authority.