When the Algorithm Became an Editor

Does journalism end when an editor clicks the “publish” button? A few years ago, this question might have seemed strange because the answer appeared self-evident: a newspaper publishes a story, and the news then reaches readers through its pages or screens. Today, however, publication is merely the beginning of a new journey—one that is no longer controlled by the media organization alone, but increasingly shared with recommendation systems and algorithms that determine which stories deserve visibility, who sees them, at what moment, and how often.

When psychologist Kurt Lewin formulated the concept of “gatekeeping” in the 1940s, and David Manning White later applied it to journalism in 1950, the gatekeeper was a person sitting inside the newsroom, deciding what would pass through to the public and what would remain unpublished. The media institution’s authority began with the editorial decision and extended to the point at which the newspaper reached the reader. The digital environment, however, has redistributed that authority without eliminating it. The question is no longer simply: Who decides whether a story gets published? It has become: Who decides whether the public sees it at all?

This, in my view, represents the most significant transformation in digital media. Power has not shifted from the editor-in-chief to the algorithm, as some suggest. Rather, it has shifted from the decision to publish to the decision of reach. Media organizations still control the production and publication of news, but they no longer have sole control over its distribution. Reaching audiences now depends increasingly on complex recommendation systems that do more than simply rank content: they determine its chances of visibility based on hundreds of signals related to users’ interests, behavior, and engagement.

This does not mean that the algorithm has become a journalist or an editor-in-chief. Such a claim would be an exaggeration unsupported by the academic literature. Recent studies, in fact, indicate that humans remain at the center of editorial decision-making. Yet today’s editors operate within an environment continuously shaped by platform logic. They know that one headline may achieve greater reach than another; that a video capable of retaining viewers for a few additional seconds may receive wider distribution; and that a story that fails to generate immediate engagement may quickly disappear, regardless of its journalistic value. The algorithm, therefore, does not dictate the editorial decision, but it reshapes the incentives under which that decision is made.

Recent scholarship supports this understanding. A 2024 study published in Journalism found that audiences place greater trust in news selected by journalists than in news selected by algorithms, while a preference for algorithmic selection did not lead to increased trust in news. This confirms that the public continues to regard human judgment as an essential element of the editorial process.

The 2025 Digital News Report, published by the Reuters Institute for the Study of Journalism, similarly shows that traditional media organizations are experiencing a continuing decline in their ability to reach audiences directly, as news consumption shifts rapidly toward social media networks and video platforms. This transformation gives recommendation systems an increasingly influential role in determining what people see, without eliminating the need for professional journalism. The report also indicates that audiences are receptive to the use of artificial intelligence to assist journalists, while remaining cautious about news produced automatically without human oversight.

Significantly, the academic debate itself has begun to evolve. The literature no longer focuses exclusively on the traditional gatekeeper. The concept of “algorithmic gatekeeping” has emerged—not to equate the algorithm with the editor, but to explain how platforms have become participants in determining what reaches the public sphere through systems of recommendation, filtering, and ranking. This process has economic, cultural, and political dimensions; it is not merely technical.

This also helps explain how newsroom practices themselves have changed. Editors no longer monitor only the quality of the text. They also track reach metrics, analyze audience data, test different headline formulations, and repackage stories for multiple platforms. The success of a news story is no longer measured solely by its accuracy or importance, but also by its ability to pass through digital gateways that the media organization itself does not control.

The greatest mistake, however, would be to assume that these metrics should drive editorial decisions. Ultimately, they measure what attracts public attention, not necessarily what serves the public interest. Journalism was not created simply to provide whatever attracts the most attention, but to provide what society needs to know—even when that information is less appealing. If digital metrics become the ultimate arbiter, journalism risks gradually abandoning its public mission in favor of the logic of competing for attention.

The real challenge facing digital media, therefore, is neither to resist algorithms nor to surrender to them. It is to redefine the relationship with them so that data remains a tool that helps journalists understand their audiences rather than an authority that determines what they should write.

Editorial independence does not begin at the moment a journalist writes a story. It begins when the editor refuses to allow the criteria of reach to become the criteria of truth. Only then can the human remain the gatekeeper—even if the gate itself has changed.