The content maze defence, books survive AI, and how Hello! escaped the no-data trap

The AI citation map, the September deadline for bots, UK publishers have a traffic surge, and Google ran out of courts - all in this week's Content Aware.

Published: 15:10, 09 July 2026

Corbidge comments on...the match report that AI can't write
The World Cup is in full swing and while our resident Sofa Strategist Rob hasn't watched a single game live, he has been reading post-match reports from two subscription UK sites. His verdict: written sport content, which comes from good journalism and combines colour, humour, and context, is still an amazing product. A piece from Greg Krehbiel, a US publishing strategist, stuck with him the most. Krehbiel argues that free AI-produced content will dominate large parts of the market because people will go for the cheaper product, even if it's of lower quality. The solution is either to go all-in on AI and run with the tech giants, or simply become impossible to erase. For Rob, the World Cup proved the second point. Ai can't see the game and it can't feed our minds the way another human can. However, will people pay for that?
Read

Publishing & Media

AP puts the US in SPUR
AI licensing standards org SPUR has its first US member, as the mighty Associated Press joins a roster of more than 30 others organisations and media houses. SPUR is developing a set of technical standards which will make it a breeze for AI firms to pay for content from creators and rights owners, building on a shared set of telemetry standards which let content businesses see how and who uses their content.
Read

Booked solid
US bookstore numbers are up 70% between 2021 and 2025, with 73% of independents reporting higher sales last year. The recipe? Part curation and part community, as readers - particularly Gen Z - who lack socialising spaces seek offline events and places to go and meet up. While shorter attention spans and other channels are hurting certain genres, people do still read and the appeal of the bookshop itself is rising.
Read

Infinite mazes for bad bots
While French title Le Monde already has licensing deals with OpenAI, Perplexity, and Meta, it is now facing an issue of bad bots outnumbering the good ones. So, it is investing in defences: bot management that spots non-human behaviour patterns, infinite content mazes that waste the time and budget of scrapers, and a payment system which requires cleared transactions before any content access. The idea is to make content scraping so hard and complicated that it's just not worth the effort. Press Gazette has the details.
Read

Local trust erodes 
Trust in local news is reducing, dropping from 82% to 70% since 2016 among US consumers, while the share who believe local news is extremely or very important to their community fell 10 points in a single year. 36% get local news from influencers, and adults under 30 prefer social media over news websites or TV for local information. While local outlets still rate higher than national outlets, the gap is narrowing as the next generations aren't paying attention in the ways the industry became used to. Nieman Lab shares more info.
Read

Hello! stopped the content guessing game
As Google referrals declined, Hello! Magazine didn't just sit and wait, they created a single daily view across Apple News, MSN, NewsBreak, SmartNews, so that their editors could see what worked where. First they noticed that a small number of stories were underperforming everywhere, then they noticed that content behaved differently across channels, with royal coverage being strong on one while lifestyle features were winning on another. After six months, page views grew between 33% and 102%, depending on the platform, and they showed that the solution isn't to find a new platform when things seem grim, but to know what your content is worth across all of them. As their CMS providers, hat tip to Hello!.
Read

Cloudflare gives crawlers a deadline 
Cloudflare announced that, starting 15 September, they will block "mixed-use" crawlers from pages that host ads by default. That means bots that blend search, agent use, and AI training into one will be axed unless site owners explicitly allow them. Besides that, Cloudflare are also evolving their Pay Per Crawl marketplace into "Pay Per Use", letting publishers charge when their content creates value in AI answers, and not just when it's fetched.
Read

UK Local News Fund: apply now
The UK Government has opened bidding for its Local News Fund, offering up to £12 million over two years to help local outlets innovate toward a sustainable digital future. Individual organisations can bid for up to £125,000, with a separate pot of up to £275,000 designated for industry-wide infrastructure projects. In their second year, £1 million is specifically for dealing with "news deserts", which would be the 37 local authority districts that have no dedicated print, online, TV, or radio outlets, affecting an estimated 4.4 million citizens. It's a part of the government's broader Local Media Action Plan, and open to outlets in England and Wales if they can prove they are delivering local news and have an audience. You have until August 7 to apply.
Read

What AI chatbots actually cite
After analysing 126 million AI prompts, Semrush found which sources chatbots reference and why. Each chatbot leans on different resources, while the brands that are cited the most are not the ones with the most content, but with the most focus, and each industry is unique. Social Media Today shares more info.
Read

A news traffic bump
According to Similarweb data, seven out of the top ten UK publishers have seen positive monthly traffic movements, with BBC remaining the dominant force and New York Times sitting static in the UK market while being one of the two publishers that see annual growth in total traffic. While the broader picture is still one of decline year on year, month-to-month we see signs that diversified channel strategies are starting to show up in the numbers - plus lots of news people want to know about. Will a World Cup bump reach extra time in future figures? Harry Clarkson-Bennett investigates.
Read

Social content in Google Search Console
Google Search Console is rolling out "platform properties" to see how your content on TikTok, Instagram, X, as well as YouTube ranks within Google Search, revealing clicks, impressions, and which search terms bring people to your social posts. Previously publishers didn't have a way to see how content from external platforms performed in Google Search. Now after verifying a connection inside the Console, you get access to performance reports, insights, and milestones. The roll out is gradual over the coming weeks and will be separate from the new search profiles features. Read more about it here.
Read

From inside Glide

Fans vs audiences unpicked
You'll never meet a football fan that stopped following a club because a rival had better content. That kind of loyalty is what gives sports organisations something that traditional publishers can't match, but can definitely learn from. At Glide Live, we had leaders from The FA, Watford FC, and Sportsbeat discussing how behind-the-scenes access took years to earn, how players' own channels are now treated as reach multipliers rather than competitors, and how one email campaign managed to generate seven figures in a single season. Sports expert Mark Gilbert writes about how sports orgs refuse to lag behind publishers, while also bringing an audience that doesn't need convincing to show up.
Read

Big Tech

Time to pay up
After seven years of appeals, the EU's highest court has upheld Google's record €4.1 billion fine for abusing its Android monopoly by bundling Search and Chrome as defaults on every phone that licenses the operating system. In 2018, Google adapted its agreements and moved on, but did they really change the default?
Read

Meta's creepy Muse skips consent
Meta's new image model, Muse, allows anyone using Meta AI to feature the likeness of any public Instagram profile (yes, you too) to create hyper realistic images, then share via chat, Stories, or feed. There is no consent question, no opt-in, no approval process, or notification to let the affected profile know. Google has a similar feature but it is limited to the user's own likeness and there is also an approval step. Meta just straight up skipped consent. This lands as Meta faces $1.4 trillion in proposed penalties from four US states suing over teen mental health harms, with trial set for August.
Read

AI & Copyright

Pot, meet kettle
Midjourney is being sued by Universal, Warner Bros, as well as Disney, for training on their content without permission. Now Midjourney is hitting back, demanding that all three reveal their internal AI usage, which includes training datasets, business plans, and board presentations. Their argument is that if the studios are doing the same thing they are suing over, then that's a fair use defence.
Read