Platforms now outrank both TV and news sites. What publishers do next is a distribution decision

The Reuters Institute's 2026 Digital News Report recorded a crossover: social media and video platforms at 54 percent as a news source, ahead of television news at 52 percent, news apps or websites at 51 percent and radio at 21 percent, across surveys of almost 100,000 people in 48 countries. Two and three points are narrow margins, and the ranking could move again. The finding underneath it will not. More than half of respondents aged 18 to 24 said social media, video networks or AI were their main source of news, and that is a formed habit rather than a phase. The number that should concern publishers most is the small one. Ten percent used AI chatbots for news weekly, up from seven percent a year earlier. In absolute terms that is minor. As a growth rate on a base that did not exist three years ago, it is the line to watch, because a chatbot is a categorically different intermediary from a feed. That distinction is the whole strategic question and it is often blurred. Search sent readers to articles. Social feeds sent readers to articles, less reliably and on worsening terms, but the article remained the destination. An answer engine removes the destination. The reporting still has to be done and paid for, and the mechanism that paid for it was the visit that no longer happens. Publishers have three available responses and they are not mutually exclusive. The first is to stop depending on the visit. Subscriptions, memberships, events and licensing all monetise a relationship rather than a page view, and they are resilient to distribution change precisely because they do not require the reader to arrive by any particular route. This is the response most large publishers have already chosen, and its limit is that it works for publications with a distinct enough proposition that someone will pay, which is a minority. The second is to be where the audience is, properly rather than as a syndication afterthought. Producing for video platforms and messaging as native formats rather than repackaging articles is expensive and requires different skills. It also concedes the economics to the platform, which is why publishers resist it and why the ones that commit tend to do better than the ones that hedge. The third is licensing to the answer engines, which several large publishers have pursued. It converts an existential threat into a revenue line, and it has an obvious weakness: it prices a publisher's archive as an input rather than its journalism as a destination, and the negotiating position weakens as models become less dependent on any single source. For the Gulf specifically, the global average understates what has already happened. Regional news consumption is mobile-first and heavily video-oriented, Arabic distribution runs through short video and messaging to a degree the aggregate figures do not capture, and platform share here has been high for years. A regional publisher planning around a television or website baseline is planning around a market that moved some time ago. The Arabic dimension of the AI channel is the one worth flagging, because it is unresolved rather than merely difficult. As chatbot use for news grows, the answers Arabic-speaking readers receive will be drawn from whatever those models were trained on and whatever they can retrieve, and Arabic coverage of regional news in general models is uneven and far less scrutinised than English. Whose reporting ends up inside those answers is being decided now, by training and retrieval choices that no publisher is party to. Work on regional models, including open-weights Arabic releases, is relevant to that in a way it is not usually framed. The measure that matters for any individual publisher is not the global chart. It is their own referral mix over the past twenty-four months, split by source, with the trend line extended. Most publishers have that data and comparatively few look at it as a strategic document rather than a monthly report. The finding to act on is not that platforms overtook television. It is that the channel growing fastest is the one that may never send a reader anywhere, and that the response has to be chosen before the trend line reaches wherever it is going.