Search referrals to publishers have fallen sharply enough that some media companies are now debating whether to block Google's crawlers entirely, according to a Wall Street Journal report summarised by Nieman Journalism Lab on 22 July 2026.
More than 85% of people worry about online disinformation, and 87% say it has already damaged their country's politics, according to a Unesco-commissioned Ipsos survey of 8,000 people in 16 countries.
Facebook's own moderator guidelines allow a post telling users to snap a woman's neck, but ban "Someone shoot Trump", Ars Technica reported in May 2017 after The Guardian published more than 100 internal training manuals.
A behind-the-scenes clip from a Palestinian short film, repackaged as proof that Israel stages fake deaths, drew about 2m views on X. It was one of several false claims that researchers and newsrooms traced back to identifiable networks rather than spontaneous user error.
Search referrals to news sites have fallen by a third globally in a year, and by 60% over two years at the smallest publishers, according to Chartbeat data. Some media companies are now weighing whether to block Google entirely.
Three hundred French newspapers have complained to the competition regulator, and a Munich court has held Google liable for the content of an AI-generated summary. In Europe the fight over visibility and copyright has moved out of the negotiating room and into courts and regulators' offices.
Reach, the British publisher of the Mirror, is cutting 220 editorial jobs and closing three local sites while creating 60 new positions. In the US, McClatchy is shedding more than 90 people, only to recruit AI editors a few weeks later. This is no longer a forecast of transformation. It is the bill.
Data from BrightEdge shows that in August 2026 ChatGPT accounted for 95.1 percent of the traffic that AI assistants referred to publishers. Gemini, Perplexity and Claude split just under 5 percent between them. The promised diversification of AI channels turned out to be a myth.
China's market for generative search optimization is set to pass 90 billion yuan in 2026 and grow 160 percent a year. The money brought practices nobody planned for: fabricated research reports that models cite as facts.
Bennett, Coleman & Co is putting its content into ChatGPT, but it will not say how much the deal is worth or how long it runs. The arrangement shows what the standard publisher-AI lab deal looks like today: content and citation in exchange for undisclosed money.
ByteDance's assistant, the most popular AI app in China, has added three paid tiers alongside a free basic version. Behind the move sits a bill nobody wants to show in public: the cost of serving 120 trillion tokens a day.
Publishers want a court to penalize OpenAI for withholding training data. The US Department of Justice argues that training models on other people's texts counts as fair use. The case is entering its decisive phase.
Since September 1, 2025, AI-generated content in China must carry labels, both openly and in metadata. Sina checked more than a dozen platforms and found three different approaches, from automatic watermarks to labels buried in the share menu.
Google pays when content "significantly contributed" to an AI answer. Cloudflare bills for use. Microsoft runs a licensing marketplace. The three differ not just over money but over who sets the terms.
A video of a protest in Warsaw, styled to look like television news footage, drew more than 200,000 views before anyone confirmed the event never happened. In the same period, the president signed a law implementing the EU's AI act.
Netflix will launch its cheaper ad-supported plan in Poland in 2027, but in May it raised prices on every tier. Poland is one of the platform's cheapest markets, yet measured against the average wage the country fares clearly worse than Germany.
Bots that handle search and AI work at once now hit a default block on pages that carry ads. Cloudflare argues most internet traffic no longer comes from people.