Media / Opinion
May 17, 2024

The value of news in the age of AI and platform regulation

Two parallel developments are reshaping the landscape of news media: generative artificial intelligence (AI) and platform regulation. These dynamics are redefining the relationship between news media and digital platforms, with significant implications for the political economy of journalism. In that sense, there is both a continuation of the “platform press” era and a divergence. This can be more clearly traced through an analysis of the strategies deployed by involved stakeholders, such as regulators and tech companies, with respect to this dual development. 

To that end, it is worth examining the strategic adjustments of key tech companies, namely Google and Meta, in response to these changes. In doing so, we can shed light on how these tech firms have framed the value of news and explore the broader implications of these strategies for the role of journalism in our societies. Such an analysis is crucial to consider alternative approaches to journalism.

The Strategies of Google and Meta

Convergences

Google’s ties with the news industry go back a long time, from the early aughts when it first inaugurated Google News following the terrorist attacks on September 11, 2001, to the Google News Initiative (GNI), its global journalism funding program, which launched in 2018. Meta (then Facebook) entered the news industry belatedly, launching its own global journalism funding program in 2017, the Meta Journalism Project (MJP). Research has shown that these programs primarily awarded funds to news media and publishers as project-based grants, training programs, advertising credits, and so forth. Platform companies carried out these programs with the help of third parties, which I call “funding intermediaries,” and which have relations to the news industry, ranging from publisher associations to education institutions.

These programs had two primary consequences: a) they were used to turbocharge the platformization of the press by making journalists and publishers all the more dependent on platforms’ infrastructures to produce, distribute, and monetize news content; and b) they served as a well-orchestrated public relations strategy for the platforms to lobby against regulatory efforts across several parts of the world that aimed to rein platform power in and, in some cases, to compel platforms to remunerate publishers for the use of their content.

These historical ties by Google – and Meta, to an extent – have not only shaped the process of platformization of news but have also laid the groundwork for current regulatory models and strategic realignments in response to new AI technologies. In other words, they established platforms as legitimate and unavoidable actors in this field. Since approximately 2022, these programs were gradually discontinued or structurally altered. The GNI is increasingly morphing into Google Showcase, a licensing program introduced back in 2020 as a “$1 billion investment in partnerships with news publishers,” while the MJP has effectively closed shop as Meta has aggressively distanced itself from journalism. A case in point is that Facebook News, Meta’s licensing program akin to Google’s Showcase, was shut down before it ever properly kicked off.

Divergences

Google: from the GNI to the Google Showcase

In the past couple of years, Google has readjusted its approach to journalism. It has done so in two ways: a) by moving away from the predominant approach of one-off grants based on so-called “innovation” projects, and b) by doubling down on licensing agreements with publishers through Showcase.

These developments cannot be examined outside of the context of generative AI and regulation. For instance, OpenAI is steadily expanding its partnerships with publishers with the likes of Axel Springer, the Associated Press, Le Monde and the Financial Times. Google is playing catch-up with Microsoft and its quasi-subsidiary OpenAI, and needs high-quality content (i.e., factual and structured data from reputable sources) to train its own large language models (LLMs) for its portfolio of commercial products, including Gemini, Generative Search Experience, and the journalism-oriented Pinpoint and Genesis. Gone are the days when crawlers freely devoured the Internet for such content, as more and more publishers are disabling scrapers’ access to their websites, raising softer or harder (pay)walls, and even suing companies such as Microsoft and OpenAI for copyright infringement.

Copyright is thus a thorny issue in this space. It is even more so in the European context, because it has effectively served as the basis for agreements for compensation between press publishers and platform firms since the adoption of the new EU Copyright Directive in 2019. Google has the necessary infrastructure to comply with the directive in a more subtle, systematic, and streamlined way. Consider Showcase, which is currently available in 25 countries. It requires participating publishers to publish three articles per day and is based on the “For You Page” model, with its feed allegedly being structured using a mix of algorithmic personalization and human curation. However, as usual, the terms of these agreements are completely opaque, with some purporting that there is also a clause requiring exclusivity for the content published there.

It’s rather clear, in my view, that Google chose this strategy because it secures a relatively steady stream of content to train its LLMs and enrich its offerings as generative AI becomes more embedded in its suite of products while insulating itself from potential copyright hassles, either in the form of lawsuits or remuneration claims. All the while it furthers the platformization of journalism by integrating AI tools into the value chain of journalism. Put differently, Google's shift towards licensing agreements appears to be a self-regulated compliance strategy with public regulatory frameworks, which allows the company to leverage its AI technologies by ensuring a steady supply of content and further entrenching the dependencies of news media on its range of infrastructures and services.

Meta: Into the Metaverse and Out of Journalism

Meta has taken a different approach than Google. Reading Mark Zuckerberg's comments from a 2019 quarterly earnings call in 2024 shows just how much has changed: “For news publishers, the thing that I'm excited about is, look, I do think that we have a responsibility to help work with news publishers to fund high-quality journalism. It's no secret that the Internet has disrupted the business model for journalism. And I think that that means that the major Internet platforms have a responsibility to form partnerships and help to fund this work.”

Fast forward to today, when Meta’s decision to block access to news for Canadian users, with negative consequences especially for smaller publishers, in response to the Online News Act (or Bill C-18) serves as a case in point that Zuckerberg has abandoned any responsibility to journalism. This behavior, naturally, neither originated in Canada nor was initiated solely by Meta. It has been part of the platforms’ playbook ever since Google blocked Google News in Spain back in 2014 in retaliation for a bill that would allow publishers to get paid based on copyrights. This tactic was also taken up by Meta and initially exercised in Australia in 2021, following the adoption of that country’s News Media Bargaining Code (NMBC), which compels Google and Meta to negotiate deals with publishers based on competition concerns regarding the duopoly’s domination of the advertising industry.

Today, it is clear that neither of these companies care about the quality of the information ecosystem or the value of journalism. The current phase of Meta’s divorce from the news industry coincided with the mushrooming of regulations, Meta’s questionable investment in the Metaverse, and the general boom of AI. Other examples of the company’s retreat from news and journalism are the termination of news-related products, like Facebook’s Instant Articles, and the decision to demote political content on Instagram. Surprisingly, the European Commission has already inquired into the company’s decision to demote political content under the Digital Services Act (DSA).

At the same time, other powerful tech companies like Microsoft, OpenAI, and Apple are expanding their reach into the news industry through commercial agreements, looking to take advantage of this critical juncture. In addition to this, the material infrastructure, cloud computing, and data infrastructures required to operate digital journalism are already captured by Google (Google Cloud), Microsoft (Microsoft Azure), and Amazon (Amazon Web Services).

Taking stock of all the above, it is clear that platforms view news either as a component to refine their LLMs and other AI-driven technologies or as a means to polish their lobbying campaigns and recover their public images after numerous backlash incidents. Engaging with news and journalism is merely a means to advance their political and economic interests. Public regulation and state power have also primarily framed the value of news in similar terms. To be clear, regulatory efforts to redirect some much-needed money to an industry that has been heavily harmed by predatory and abusive business practices of a tech oligopoly is the just thing to do. Anya Schiffrin and other colleagues have done important work on this, in particular concerning the appropriate economic value of news to tech firms in the context of regulatory frameworks like the NMBC. Yet, I believe that we should expand our scope of understanding of the value of news to embrace more collectivist frameworks, which are not confined to a purely economic understanding.

In Need of a Different Vision

Recently, I came across an op-ed by the director of Médiapart, Carine Fouteau, a beacon of independent and engaged journalism in France, in which she claimed that her organization will not touch Google’s money so long as it comes with strict non-disclosure agreements. Without delving too much into the context, the gist of the story is that some French publishers entered into negotiations with Google and Meta to ensure remuneration for the use of their content based on the EU’s Copyright Directive. However, as we’ve written elsewhere, Google abused its power and attempted to bypass the spirit of the law and trick publishers into unfavorable terms such as combining the license for Showcase with the copyright license. Thankfully, the French Competition Authority stepped in – twice – and sanctioned Google, demanding that the company comply with specific measures when negotiating with publishers.

At any rate, some smaller and independent publishers were left out of the discussions and deals. As a result, a broad coalition of other publishers came together to create the Société des Droits Voisins de la Presse (DVP), a Collective Management Organization, to negotiate collectively with platforms, and to collect and distribute funds equitably to its members. Médiapart supported this initiative because the only viable way of dealing with such corporate power is through collective action. Such divide-and-conquer tactics, reinforced by information asymmetries both between platforms and publishers and among the publishers themselves, have been long employed by Google and Meta. Now, they are also being employed by new entrants like OpenAI. Collective action can mitigate the harms of these tactics.

The French case study holds another important lesson: the law can only do so much. Google never really complied with those requirements because it could sustain the financial blow of the hefty fines (approximately $800 million). It requires another approach and to practice resistance, as evidenced by Médiapart’s example, which comes in direct contrast to what other French – and international – publishers have been doing: going at it alone and striking deals with tech companies at the expense of the collective values of the community. This could mean creating publicly accountable funds for news media, supporting collective management and bargaining efforts, facilitating more non-commercial models, and the broader decommodification and democratization of the relations of journalistic production. Even bolder options should be on the table, like more community- or worker-owned models and supporting public service media. Crucially, our regulatory frameworks should accompany and empower such models and acts of resistance.

As a result, I think what is already changing – and is bound to change even more profoundly – is the value of news, or better put, the value that is assigned to news (and journalism) by different stakeholders. It seems that some policymakers understand this. For instance, the EU’s European Media Freedom Act (EMFA), which recently came into force, contains an article that assigns a must-carry quality to news circulating on platforms to protect it from the arbitrariness of platform content moderation mechanisms. Despite its many problems, this regulation inherently ascribes a democratic value to news beyond a narrow understanding of it as another type of content or commodity. If we want to preserve journalism’s normative (albeit often idealized) role in a functioning, respectful, and inclusive democracy, then we, as a society, should be striving to assign it a value that is conducive to principles of collectivism and communal flourishing rather than corporate and individualistic ones.

We now reach the core of what’s at stake: how we perceive the value of news reflects how we perceive the role of journalism in and for our democracies and, by extension, the structures of our societies. Meta’s abrupt and hurtful divorce from the industry has left a platform-dependent news industry scratching the barrel for ways to make up for the loss of traffic and financial support. All the while, Google is expanding its Showcase program to consolidate, protect, and futureproof its market power.

Another vision is required. In the age of AI, the value of news emerges – if there was ever a doubt – as another site of struggle and contestation in which we should collectively participate to challenge bankrupt neoliberal ideas and models.

This article was first published on Tech Policy Press on May 16, 2024.

More results /

/ algosoc
Emerging generational fault lines in public opinion on Artificial Intelligence

By Ernesto de León • Fabio Votta • Theo Araujo • Claes de Vreese • May 17, 2024

Sustainability in the digital world

By José van Dijck • May 06, 2024

/ justice
Artificial Intelligence and the Judiciary

By Corien Prins • November 21, 2023

Rutte IV: toezichtreflex en Autoriteit Persoonsgegevens

By Corien Prins • January 28, 2022

/ health
Discriminerende algoritmes

By Corien Prins • May 18, 2021

/ media
Content moderation and platform observability in the Digital Services Act

By Charis Papaevangelou • Fabio Votta • May 29, 2024

Journalism in the age of AI, platformisation, and regulation

By Agustin Ferrari Braun • Charis Papaevangelou • May 27, 2024

The value of news in the age of AI and platform regulation

By Charis Papaevangelou • May 17, 2024

Subscribe to our newsletter and receive the latest research results, blogs and news directly in your mailbox.

Subscribe