Skip to main content

Blog

Bargaining in the shadow of the press publishers’ right: A new working paper

Posted on    by Ula Furgal
BlogDealing with creatorsWorking papers

Bargaining in the shadow of the press publishers’ right: A new working paper

By 10 April 2024July 9th, 2024No Comments

CREATe is happy to present the fourth entry in our series of working papers released in 2024: ‘Bargaining in the Shadow of the Press Publishers’ Right’ by CREATe’s Ula Furgał and Martin Kretschmer. The paper is linked to the ‘Rethinking News and Media Law in Europe’ project of the Institute of Information Law, University of Amsterdam, bringing together copyright, human rights and media scholars to debate an integrated new approach to European media law. The paper, alongside other project contributions, is forthcoming in the Cambridge Handbook of Media Law and Policy in Europe (CUP 2024).

It has been five years since the adoption of the Copyright in the Digital Single Market Directive (CDSM). Still, the press publishers’ right it introduces, continues to generate controversy. Following enactment of the new right, the publishers’ exclusive right to their news content is no longer disputed. However, how it is supposed to translate into remuneration from digital platforms remains a puzzle. Our paper addresses this question.

The CDSM Directive provides little to no guidance on how the licensing negotiations between press publishers and platforms should proceed. The expectations that money will flow only because there is a new related right were rather misled, as Google’s refusal to negotiate in France in 2019 have quickly shown. Inspired by the Australian News Media Bargaining Code, EU member states started to fill this gap in their (much delayed) national implementations, shaping the negotiation frameworks to facilitate, or sometimes to force, bargaining between press publishers and platforms.

The paper we are publishing today, analyses those negotiation frameworks through the lens of the literature on ‘bargaining in the shadow of the law’ to draw wider lessons on the relationship between statutory law and private bargaining. It identifies four main approaches to the operationalisation of the press publishers’ right in the EU: 1) competition law commitments offered by Google and approved by the French Competition Authority (AdLC); 2) collective management and licensing with extended effect in Denmark; 3) final offer arbitration and the key role played by a regulator in Italy; 4) Extended News Previews (ENP) program established by Google; and constructs those frameworks alongside the Australian News Media Bargaining Code, a mandatory bargaining scheme not grounded in an IP right.

Click on image to download

By defining and coding key parameters concerning the bargaining parties, their endowments (bargaining chips), measures to mitigate dependencies and the role played by the authority (such as a regulator), the paper shows that bargaining frameworks sit as complex institutional constraints between pure private ordering and the law. It is not possible to predict the shape and the frequency of the agreements struck by publishers and digital platforms solely by looking at the press publishers’ right wording. It is the bargaining framework which is the key. That said, the agreements made within all the frameworks we analyse are not available to the public, signalling a broader transparency problem: with the press publishers’ right set to contribute to the sustainability of the press sector, recognising its role in the modern democratic societies, the public has little information on the deals made in its name.

The question of ‘how’ to operationalise the press publishers’ right might be gaining new momentum in the age of AI. In its recent decision imposing a 250 million euro fine on Google for the lack of compliance with the competition law commitments, the French Competition Authority (AdLC) created a link between the press publishers’ right and generative AI models. While the decision focuses on the calculation of the publishers’ remuneration, the AdLC also notes that by not informing the publishers if and how their content had been used in training and operating Gemini (known earlier as Bard), Google’s generative AI, Google breached its commitment to transparency. The AdLC recognised that the question of whether Gemini uses press content in a way covered by the press publishers’ right remains open, so no definitive answer as to whether publishers are owed remuneration from generative AI developers was provided.

The press publishers’ attitudes towards their content being used to train and support the running of generative AI vary considerably. While some are eager to license their content to OpenAI and Microsoft, owners of ChatGPT, granting access not only to their archives, but also to current reporting (see agreements signed with the Associated Press in the US, Axel Springer in Germany, Le Monde in France and Prisa Media in Spain), others decide to block generative AI bots from crawling their content and refuse (for example Reach, a major UK media group) to engage in licensing negotiations. Other publishers such as Semafor, develop more complex relationships with AI companies. Signals, Semafor’s news feed, uses a ChatGPT-based chatbot to offer links and summaries of articles from other outlets to provide context to Semafor’s reporting (not unlike a news aggregator), with development being sponsored by Microsoft. There are also those publishers who choose the litigation path: while The New York Times was first to sue OpenAI and Microsoft in December 2023 for copyright infringement, they were recently followed by smaller news outlets: The Intercept, Raw Story and AlterNet in the US, taking the route of infringement under the Digital Millenium Copyright Act (DMCA) instead.

While the press publishers are more hesitant to feed their content into generative AI for free than they were to have it included in search results and share it on social media, OpenAI’s and Microsoft’s picking and choosing whose content to license carries a strong resemblance to the deals signed earlier by Google and Facebook for their services.

Our analysis does not imply that payments from AI companies should be mandated by law, and become central to the negotiation frameworks developed under the press publishers’ right. Rather, we urge caution. Payments from big tech should come with a dependency warning: only when a press publisher’s content is used, the publisher will be remunerated, potentially creating a vicious circle where press publishers become dependent on income from a handful of big players. This may make it more difficult to develop income models controlled by the publishers themselves. In order to ensure quality and pluralism of news, the press needs to break free.

Bargaining in the Shadow of the Press Publishers’ Right

Ula Furgał and Martin Kretschmer

CREATe Working Paper 2024/04

Abstract

The press publishers’ right (granted under Art. 15 of the EU’s 2019 Directive on Copyright in the Digital Single Market) equips publishers with a legal basis for negotiations with digital intermediaries. However, the process of bargaining has not been specified by the EU legislator, who left it at the discretion of EU Member States and the market. This paper constructs the four main approaches to the operationalisation of the press publishers’ right in the EU, in contrast with the bargaining framework that evolved in Australia without the underpinning of a new intellectual property right. It analyses those five frameworks in the light of the ‘bargaining in the shadow of the law’ perspective, to understand the relationship between statutory law and private bargaining as mediated by institutional frameworks. By defining and coding key parameters concerning the bargaining parties, their endowments (bargaining chips), measures to mitigate dependencies and the role played by the authority (such as a regulator), the paper shows that bargaining frameworks sit as complex institutional constraints between pure private ordering and the law. It is the shape of the bargaining framework which matters for the range and the frequency of the agreements between press publishers and digital platforms.

Full paper can be downloaded here.