Trust makes a comeback, the bundle that keeps paying off, and the classroom catch

Brussels hits pause on social media bans, follow the data trail, Grok gets graded, and Spotify plays both sides - all in this week's Content Aware.

by Emina Zulic

Published: 15:01, 05 November 2026

Corbidge comments on... Reddit raises the drawbridge
Reddit was an outlier by being a free forum which kept on foruming at an enormous scale, driven by freely proffered user generated content. When it signed AI deals with OpenAI and Google in 2024, it hoped its vast archive of human conversation would be a ballooning asset, but two years later, with Google traffic going down the drain by more than a million referrals a week, it's now moving behind a login wall. Our walking UGC machine Rob hears the sound of a content drawbridge being raised as Reddit looks a lot like the publishers which its forums so frequently "homage" by posting acres of their content.
Read

Publishing & Media

Trust on the up
According to a new Pew Research Center survey, Americans are more likely to trust information from news organisations than they were a year ago. In September, 62% of US adults said they had at least some trust in national news organisations, up 6 percentage points on 2025, although still lower than a decade ago. Local news fares better, with 75% trusting local outlets. When it comes to social media, it is flat at 38%. The increase spans both political parties and most age groups. For publishers, local news remains the most trusted part of the industry, with social media a long way behind. 
Read

The accidental friend
While AI may be squeezing publishers now, it could end up being their best advert. Once a convincing deepfake actually causes some real harm, people will happily pay for news that they can trust, according to a former BBC journalist turned PR boss. He points out that chatbots only sound confident because they're quoting journalism, which basically makes Silicon Valley's business model "an advert for the local paper", and expects short-term pain before long-term gain.
Read

The New York Times story
While many large media brands would love to have 420,000 paying subscribers in total, the juggernaut New York Times added another 420,000 digital subscribers in the last quarter alone, taking its total to 13.8 million and on track for 15 million by the end of 2027. Revenue rose by almost 10%, digital subscription revenue was up nearly 15%, and digital advertising grew 12% even as print subscribers continued to fall. While much in the industry fight over diminishing referral traffic, the Times shows that good content well delivered to people willing to pay for it is still the fundamental business model of media. 
Read

One switch to rule them all
Google, Bing, OpenAI, and Anthropic have all said they'll support a shared robots.txt control being developed by the IETF, the body which sets internet standards. It is a major boost in efforts to set common standards which would allow site owners to more easily manage bot requests to their sites. It would allow site owners to opt out of AI training across every participating provider in a single step, instead of dealing with each company's crawler separately. While it is welcomed, Google won't support it until 2027, and it only covers LLM training, not whether your content appears in AI answers such as AI Overviews, so think of this more as a useful step towards the full control site many are campaigning for. 
Read

Gemini's copyright problem
In the first generative AI case to reach the EU's top court, an adviser has said that Google's Gemini may infringe publishers' rights when it retrieves their articles and reuses them to answer user questions. That applies in cases where it gets around paywalls or bot blockers, or reuses the article's wording in a recognisable way that it substitutes for the original. If Gemini wants to train on the same articles, that's currently permitted under EU law. The court usually follows its advisers, and if a ruling comes in the coming months it could give publishers a legal route against AI answers that replace their reporting, even if it can't do much about the secondary background training for now.
Read

Wikipedia's AI dilemma
While Wikipedia's own boss calls AI an existential threat, a journalism professor argues that even Grokipedia, the sometimes-correct AI rival from the spaghetti of Elon Musk's portfolio, has something to teach it: AI is good at digging up details that humans would miss, as long as people are the ones who decide what's true and what matters.
Read

Google's hidden dials
An SEO agency has published a list of 120 "twiddlers", which are the programs Google uses to rework search results after its main ranking, boosting some pages, demoting others, and spacing out results from the same site. It offers a rare glimpse into what Google's engineers are working on: a twiddler for E-E-A-T expertise judged page by page, a filter still named after Gemini's predecessor Bard, and a site-wide demotion once a site racks up enough copyright takedown notices. Scour the full list here.
Read

Meta's hostage situation
Publishers united against Meta's new scraping terms, but may have less leverage than they'd like. Meta's platforms now send barely 3.5% of publisher traffic, yet heads of audience say Instagram and Facebook are still important for brand awareness, community, and even new subscriptions, so walking away isn't as simple as the percentage might imply. Trade bodies want Meta to separate access to its platforms from permission to scrape, and to write protections such as robots.txt into binding terms rather than bendy-penned assurances. If the terms don't change, some media houses may decide the trade-off just isn't worth it anymore. 
Read

Brave or foolish?
The New York Times, the Daily News, the Chicago Tribune, and the Orlando Sentinel have sued Brave, alleging the privacy-focused browser company scraped their journalism with crawlers publishers could neither identify nor block. After scraping, the suit says Brave sold access to the results to AI companies including Anthropic, Perplexity, Mistral, xAI, and Cohere, and even helped its users get around paywalls. The case goes after the data pipelines which are feeding AI search, and compares Brave to both the burglar and the fence. 
Read

From inside Glide

No more stale stories
An update is only as rapid as the last system to pick it up. With Glide Transmit, every change made in Glide CMS reaches your caches, search, apps, and partner feeds within moments, with no need for scheduled jobs and nothing to build inside the CMS for each new system you connect. Editors stay in control of where each story goes, and a delivery history shows exactly what and when each channel received in an update. 
Read

Big Tech

Bano a bano talks
The European Commission has told EU countries to pause new national social media bans for children as it works on its own 27-state-wide single EU law. The pause, for 12 months, freezes Austria's imminent plan to keep under-14s off social media, prompting Vienna to criticise Brussels for blocking national efforts before a Europe-wide alternative is ready. With the EU Kids Act not likely to take effect before 2028, there may be a flurry of national laws hitting statutes this time next year as patience for a single law wears thin.
Read

Google's school bully
Worried about AI in your child's education? Google is letting schools opt out of its AI tools in the classroom, but not without a catch. If they block Gemini AI (as many parent likely do), they also lose Google Search, which Google now counts as an AI product. Schools have the option to either allow all of Gemini's classroom tools and Search, or lose all of it. This isn't anything new from Google, which has already bundled its AI into a single opt-out for publishers, and is betting schools flooded with Chromebooks will have to bite the AI bullet. 
Read

AI & Copyright

Spotify's AI cover band
Spotify is planning to launch a paid AI remix service early next year, allowing subscribers to turn any song genre into another, or have one artist "cover" another. It has already signed deals with Universal Music Group and others, royalties would go back to the original artists, and musicians would have the option to veto the use of their songs. Some of today's biggest stars are already pushing back, and plenty feel the streaming service will end up flooded with AI remixes crowding out the actual original tracks, and new music. It's an interesting turn for a company that deleted 75 million AI spam tracks last year. Now they know how to monetise them, expect AI DJs presenting Slop 10s before long.
Read

When AI filled the gap
When fan fiction archive AO3 went offline after a security incident, a copycat site emerged like a saviour knight to "save" its nearly 19 million works. But behind it, a teenage developer had used AI agents to scrape the stories for AI-generated audiobooks to be put under paid access, with zero consent from the authors. They responded by firing off around 2,000 takedown notices in a single day, forcing a rapid backpedalling and removal of the site, an apology, and insight that it sold only 11 audiobooks. While it shows how quickly AI tools can turn someone else's work into a product, it does too that rapid vocal reaction works. 
Read

ChatGPT ad flop
ChatGPT in-chat ads aren't paying their way yet, according to a new working paper which tracked nearly 1,900 online shops. In the US, a visit from a user who clicked a ChatGPT ad brought in on average $0.89 in revenue, against $2.45 for a paid Google search visit. Perhaps they are all broke buying RAM? According to the researchers, that makes a click worth about $0.26, while one of the largest known advertisers paid $1.72 per click. The ads currently bring in about $1 per free user a year, while OpenAI is reportedly banking on ad revenue bringing about $40 per free user by 2030. With ChatGPT now claiming around 20% of Google's search volume but only a fraction of its users paying, that's quite a gap to close. 
Read