Il Washington Post has just announced that it will replace some of its editors with Ember, an artificial intelligence. This is not a joke, nor the umpteenth whim of Jeff Bezos. AI-assisted writing is the first large-scale experiment by a mainstream publication that entrusts editorial oversight to an algorithm.
The stated goal is to attract “non-professional writers” and turn them into contributors through automated suggestions and analysis of narrative structure. While some newsrooms are experimenting with AI to automate content, the Washington Post is going a step further: automates the training of journalists themselves.
Ember: The Editor That Never Sleeps
The internal project called “Ripple” has already passed the research and development phase, which began over a year ago. Lips Oosterhof, strategic consultant to the Washington Post and developer of the system, has created what may be the first artificial editor in the history of journalism. Ember does more than just correct grammar and syntax: analyzes narrative structure, suggests improvements, and guides aspiring writers through a personalized coaching process.
The “narrative strength tracker” is the heart of the system. The interface shows you in real time how your story is developing, with a sidebar highlighting key elements: “opening thesis,” “supporting points,” and “memorable ending.” A live AI assistant provides development questions and suggestions for adding “strong supporting points.” It’s like having a chief editor constantly whispering in your ear, but it never gets tired and costs a lot less.
The trial will begin this summer with editorial partnerships, then expand in the fall to testing with “non-professional writers.” The Washington Post has already identified potential collaborators: fromAtlanta Journal-Constitution al Salt Lake Tribune, passing through popular Substack authors like Matt Yglesias e The Dispatch.
Washington Post: When Business Meets Forced Innovation
The move comes at a time of deep crisis for the Washington Post. Jeff Bezos is desperately trying to turn around a company that has lost hundreds of thousands of subscribers and millions of dollars. The pressure to cut costs and find new sources of revenue pushed management towards increasingly creative and controversial solutions.
The Ripple Project aims to potentially reach 38 million American adults, according to internal research. The idea is simple: if you can't afford professional journalists, create a system that turns anyone into a competent journalist. It is the logic of the platform economy applied to traditional journalism. Will it work?
Assisted writing that divides editorial teams
Not everyone in the industry is happy with this development. As we have already analyzed, artificial intelligence in journalism is nothing new. Already in 2017, the Washington Post had launched Heliograph, its first robot-journalist to cover sports events and election results. Ember now represents a qualitative leap: it does not produce content, but trains those who produce it.
The issue goes beyond simple automation. According to the Order of Journalists, 85% of industry players expect AI to become essential for personalizing content and speeding up workflows. However, entrusting editorial supervision to an algorithm raises new ethical and professional questions.
Washington Post, the paradox of democratization
The Washington Post claims to be “democratizing” journalism by making quality writing accessible to the untrained. In reality, it is creating a system in which human expertise is replaced by predefined prompts and statistical analysis. Is it really democratization or (pardon the term) is it the definitive “commoditization” of the journalistic profession?
Assisted writing promises to free “real” journalists from repetitive tasks, allowing them to focus on in-depth analysis and investigation. But what happens when AI becomes so good that even the human oversight the Washington Post promises to maintain is no longer necessary?
Towards a journalism without journalists
The Ember project is a test case for the entire industry. If it works, every publisher will be tempted to replicate it. If it fails, it will demonstrate that some aspects of journalism remain irreducibly human. The Washington Post is toying with the very future of the profession, and the stakes are much higher than just company balance sheets.
Perhaps the real question is not whether AI can write as well as a journalist, but whether we are truly ready to live in a world where the distinction between professional and amateur content dissolves completely in the algorithm.
The Washington Post has chosen: and you, which side are you on?
