The rise of generative artificial intelligence is changing the relationship between journalism and audiences, raising a basic question: are we writing for readers or for machines? AI can provide tailored answers in seconds, so users no longer need to browse result lists or compare multiple sources, while the barriers to becoming a content creator have fallen — almost anyone can now produce publishable text.
How information consumption is shifting
A Reuters Institute study published on 29 October 2026 found that weekly use of generative AI systems across six examined countries rose from 34 to 51 percent within a year, and weekly use of AI for information gathering increased from 24 to 38 percent. In a nine-country survey, 64 percent of those using AI for information cited faster and easier answers as the reason for choosing the technology, while only 21 percent said they used it because it provided more reliable information.
These figures point to a notable change: convenience and speed are becoming dominant considerations in how people consume information. For many users, a seemingly adequate summary is enough — regardless of how many journalists, researchers or newsrooms produced the underlying work.
Economic impact and the rise of "invisible" journalism
A 2025 Pew Research Center study highlights the economic effects of AI aggregation: when an AI-generated summary appeared in Google search results, only 8 percent of users clicked through to a traditional search result, compared with 15 percent when no AI summary was present. In Reuters Institute forecasts published in early 2026, publishing executives expected an average 43 percent decline in search-driven traffic over the next three years. Chartbeat data cited in the analysis showed organic Google search traffic to the examined news sites fell by 33 percent between November 2024 and November 2025.
This creates a paradox for publishers: producing original information remains costly — journalists, experts, research, editing and fact-checking are required — yet finished content increasingly appears as part of other platforms’ AI-generated answers, and publishers risk losing the visits that historically funded their operations. The link between source and reader may weaken: if key findings of a report are absorbed into AI responses, users often never learn the author, the outlet, or the original argument.
Anyone can become an author — risks of mass-produced content
Generative AI also lowers the barriers for creating content: with well-crafted prompts anyone can generate text that looks like a news article, a professional analysis, or even an academic-style paper, often without journalistic experience or deep subject knowledge. This echoes earlier waves on social media where many people presented themselves as nutrition experts, coaches or life advisors.
The academic world is especially vulnerable. Nature’s 2026 analysis of 2.5 million biomedical studies identified nearly 3,000 publications that cited non-existent research — not all attributable to AI, but illustrating the fragility of scholarly communication. Journal editorial teams face growing verification burdens: rising submission volumes, peer review abuses, and the difficulty of detecting machine-generated writing all strain editorial processes.
International medical journal editorial guidelines already require clear disclosure of AI use and affirm that AI cannot be listed as an author: humans remain responsible for accuracy, originality and professional credibility.
AI can sound authoritative — implications for professional media
AI-generated copy competes for the same readers, search results and algorithmic recommendations as carefully produced journalism, even though editorial checks and professional judgment can be bypassed in automated workflows. There is also pressure to optimize content not only for human readers and search engines, but for AI systems themselves: if publishers prioritize machine-readability, unique voices and complex argumentation may be marginalized, yielding more uniform, easily processed pieces.
Worsening the problem, AI-produced materials can be fed back into other AI systems to create summaries and derivative texts. Each processing stage increases the risk of propagating inaccuracies, and the original source and authorial responsibility become harder to trace.
Reuters Institute research from 2026 shows that 67 percent of news consumers prefer news produced entirely by human journalists, while only 12 percent welcome content generated solely by AI. This suggests enduring demand for genuine journalistic craftsmanship, particularly in specialist media.
What advantages can newsrooms leverage?
Exclusives, original investigations, expert commentary, on-the-ground reporting and independent analysis create value that is harder to replicate algorithmically. Personal professional networks, community building and a trusted editorial brand can become meaningful competitive advantages against easily recyclable content.
The stakes go beyond pageviews and revenues: if AI systems increasingly determine which information is selected, summarized and interpreted, they will directly influence which ideas and perspectives reach the public sphere. Simultaneously, distinguishing verified, trustworthy information from surface-level summaries will become harder.
Journalism at a crossroads
I believe there is still value in the human voice: a well-asked question, a substantive conversation, or an article worth ten minutes of reading. At the same time we must confront the contradiction that while journalists strive to create material worth reading, AI development increasingly promises readers they may not have to read at all.
So who are we writing for? Today the answer seems to be both: people and machines. We must make our content interpretable by algorithms because they often mediate the reader’s path, but we must not forget that journalism’s purpose is grounded in human curiosity, professional responsibility and independent thinking — qualities no SEO trick can replace.
Personally, I prefer to write for people: those who care about the author’s identity, expertise and viewpoint, and who will sometimes read a piece from its first sentence to its last. If you have read this far, at least once I have directly reached the audience I intended.
