Publishers dealt a winning hand, TV news lose the remote, and death by a thousand edits

The Gray Lady draws the line, Canada bucks the trend, Google's "freeloader" accusation, and AI ate a quarter of your work - all in this week's Content Aware.

Published: 15:17, 14 May 2026

Corbidge comments on… the Crown Jewels under the hood
The EU's Court of Justice ruled this week that platforms must pay publishers for using their content, and that any negotiation means access to platform data. The case stems from Italy's AGCOM, and the ruling bars platforms from retaliating with reduced visibility during the negotiations. For publishers who've been told for years that their content is more or less worthless, while platforms have accumulated unfathomable wealth using that same content, this is the first real chance to see what the deal is actually worth. Our resident Revenue Ranger Rob discusses value of content, a ruling that might crack the door open, as well as that whatever is revealed in Italy, will be revealing for us all.
Read

Publishing & Media

Pay up for news
The Court of Justice of the European Union has sided with the Italian watchdog AGCOM and ruled that Meta has to compensate publishers for using snippets of their articles. The ruling forces the platforms to hand over the data that is needed to calculate the compensations, bans them from burying news content in search results during negotiations, and also gives national regulators space to step in and decide terms if the talks stall. Publishers can, if they want to, refuse the use of their content entirely. The European Publishers Council called this a massive victory and urged all EU member states to look at this ruling as a blueprint. For publishers which have been fighting these battles, the decision slides a few aces up their sleeves.
Read

Canada's quiet newsroom boom
While others are thinning their newsrooms, Canada's Globe and Mail are adding roles, 30 in the last year, which is their biggest hiring push yet. They have a simple logic: cover what people are actually interested in, not what journalists think they are interested in. While business news grabs subscribers - sport, real estate and regional stories are what keeps them paying. The paper is also pushing visual journalism hard, including even sending a reporter on a three-week trip to the Arctic, which resulted in breaking three cameras. When it comes to big tech, their editor-in-chief was clear: it's not a journalism problem, but a commercial one. Subscriptions now make up two-thirds of the revenue, print ads less than 15%, which prompts them to describe their model as "very strong". Press Gazette shares more details.
Read

No robots, please
After struggling with accusations of AI leaking into content, The New York Times has sent freelancers a reminder which bans the use of generative AI to create, draft, edit, improve, or rephrase any of the submitted work. The policy is strict and throws AI-powered search and image generators out the window. High-level brainstorming is allowed but using AI to create or edit content is a big no-no. While NYT claims this is just routine procedure, some recent public embarrassments might suggest otherwise.
Read

AS Roma and Glide: a winning team
When AS Roma set itself a bold new vision for a completely transformed digital experience for fans, with advanced new revenue management options, and faster site and apps, Glide CMS was delighted to be chosen to deliver the project. This meant migrating over 36,000 articles and 100,000 images, integrating ticketing, multiple CRMs, a data lake, online shop, and SSO into one platform, and delivering a redesigned fan experience worthy of the Giallorossi - within a tight immovable timeline. The editorial and commercial teams now spend less time in the tech and more time creating what fans want, and new ideas can be enacted far quicker without waiting for old-fashioned development cycles.
Read

Publishers build their own rulebook
Mediahuis, the publisher of Irish Independent, is the latest addition to the SPUR coalition (Standards for Publisher Usage Rights), which is a non-profit working to create shared standards for how AI platforms use journalistic content. Among the founding members we have BBC, Financial Times, Guardian Media Group, Sky News, and Telegraph Media Group, their aim being to track how content is used, establish licensing frameworks, and build measurement tools for publishers to negotiate equal terms for their content. Since their launch in March, SPUR has been laying the technical groundwork on content tracking and is accepting new member whose primary activity is original journalism. Whether AI companies come willingly to negotiate or they are dragged to the table by regulation, having publishers in a stronger position is always a good thing in our book.
Read

Same playbook, different wrapper
While AI content tools are useful, misuse to generate entire sites is a losing bet. A study tracking 220+ sites found that the brands still growing after recent Google updates are the ones whose content doesn't match the cookie-cutter templates being sold as "AI-first SEO" or "GEO-optimized content at scale." While AI tools work well for research, briefs, pulling in data and speeding up workflows - with human oversight - they're backfiring when the goal is volume and not value, or when the human oversight is missing. Lily Ray shares more details on the data and what to watch out for.
Read

The cognitive tax of AI
Our brains are all playing Sherlock Holmes nowadays, asking if things we see are AI or genuine human thoughts. It doesn't help that the researchers at Imperial College London found that 35% of new websites are AI generated, online writing is getting more bland, and that original viewpoints are shrinking every day. To make things worse, humans are either absorbing AI patterns without even knowing it, or they're being accused of using AI when they aren't. 404 Media shares more details and explains why all of this is driving everyone nuts.
Read

The broadcast frog in boiling water
Former BBC News boss Deborah Turness warns that traditional TV news is facing a bigger threat from personality-led journalism on TikTok, YouTube, and Substack, with a potentially even bigger shift than with the arrival of social media. She has pointed out the major drop in TV news audiences - nearly 4 million in five years - while the consumption of YouTube and TikTok news has tripled and grown tenfold, respectively. Turness has one clear message to TV broadcasters: shape your decisions by where the consumer is and liberate your talent, instead of making decisions with a broadcast-first mindset. Whether publishers can adapt fast enough, or if they will be left in the dust, that remains to be seen.
Read

Google points the finger, publishers point it back
In its latest submission to the UK's Competition and Markets Authority, Google has proclaimed a publisher's "free rider's charter", objecting to proposed rules that allow publishers to decide whether their content ends up in AI Overviews. Naturally, this has triggered fury from the News Media Association, which represents titles such as The Telegraph, The Times, and The Guardian, reminding everyone that Google has been syphoning value from journalism for years without any consent or compensation. The NMA has written to the CMA's chief executive warning that this is Google's attempt to "fatally weaken" the regulatory regime before it even kicks off.
Read

Big Tech

Brussels locks the server rooms
The European Commission is preparing their "Tech Sovereignty Package", which could restrict US cloud giants from processing sensitive public sector data, such as financial, judicial, and health records. The move is designed to bootstrap sovereign cloud offerings and give EU-based providers a chance to compete. Private companies are able to continue to place their data where they choose, but public procurement rules would be adapted to promote more diverse choices between cloud and AI providers. A Commission spokesperson told CNBC that "Europe was waking up and getting its act together". These new rules are expected later this month, alongside the Cloud and AI Development Act as well as Chips Act 2.0. Combined with a potential EU-wide social media ban for under-15s, expected this summer, EU is seemingly showing digital sovereignty teeth.
Read

Scam ads: a feature, not a bug
Meta has another lawsuit coming its way, this time from Santa Clara County, who is suing them in California state court. The lawsuit alleges that the company earns up to $7bn a year from scam advertising both on Facebook and Instagram. According to it Meta has "largely tolerated" fraud, set up internal guardrails to block scam-reduction efforts that were too costly, and even targeted scam ads at users who previously fell for similar fraudulent listings. This isn't the first time Meta has been connected to scams, as Reuters previously reported on documents suggesting Meta projected more than 10% of its 2024 revenue would be from scams and banned goods. While Meta claims they take enforcement seriously, these figures suggest that revenue seems to be a bigger priority.
Read

AI & Copyright

Usage up, trust down
According to Morning Consult's latest tracking data, Americans are finding more reason to use AI but at the same time are souring on the industry behind it. AI now falls under one of the least trusted sectors the firm tracks, even if people are using it more. Globally it's a bit different, as trust in AI is either stable or rising. The full report  benchmarks AI brand trust against 200 categories and 3,000+ brands across 28 markets, digging into which demographics are most sceptical as well as which institutions people trust to use AI responsibly. If you plan to build anything AI or sell AI products to English-speaking audiences, it might be worth paying attention to the gap between utility and trust.
Read

Google extends the olive branch
After two years of AI Overviews sabotage, Google is rolling out changes that will add more publisher links inside their AI answers, with features such as "Further Exploration" linking to relevant articles, "Expert Advice" showing snippets from news, and more source notes in general. Google is also testing subscription integration, which would link the users existing subscription with their Google account so paywalled sites are surfacing more in AI results. According to the early tests, users are more likely to click through if the subscribed sites appear. With everything going on around Google, from lawsuits to the DMA in full effect, this doesn't feel like generosity but course correction before more trouble arises.
Read

AI's 25% content tax
According to Microsoft researchers, frontier AI models lose an average of 25% of document content over 20 interactions, with weaker models getting up to 50% of degradation. They tested 52 professional domains, from accounting to music notation, and found only one where models were "ready": Python programming. In 80% of model/domain combinations, the results were "catastrophic corruption" and to make things worse, giving them tools caused more degradation. Thanks to errors happening all at once, a single interaction can wipe 10 to 30 points. To make things worse, according to another Microsoft study, AI agents can be reliably tricked into making bad decisions by tactics that no human would fall for, with even frontier models showing vulnerability.
Read