One date to rule them all, journalism strike season, and Europe turns up the heat on Google

The mouse breaks up with OpenAI, AI skips breaking news, music industry's bad day at court, and headline hijack - all in this week's Content Aware.

Published: 15:53, 26 March 2026

Corbidge comments on…when the platforms finally face the music
Two significant legal verdicts landed this week. Meta was hit with a $375m penalty in New Mexico for failing to protect children from online predators, and a Los Angeles court found both Meta and Google liable for the depression and anxiety suffered by a young woman addicted to their apps. The cases weren't about the content on these platforms, but about the mechanisms put in place to keep users glued, and the consequences which inevitably followed. The Corbidge Correspondent shares his thoughts.
Read

Publishing & Media

Journalists on picket lines
It's a bad week to be a media manager. In Australia, ABC journalists and technical staff walked out for 24 hours yesterday, their first strike in 20 years, after 60% of the workforce rejected a below-inflation pay deal. On the other side of the world, in New York, ProPublica's unionised staff voted 92% in favour of the same action after more than two years of bargaining for a contract. Two very different newsrooms, one underlying frustration which both creative teams and media business owners can agree on: content isn't cheap, and making money from it is hard. Will the social media and Google adtech trials help swing the needle?
Read

Old date, new problem
Here comes a quick but useful SEO reminder: if your pages show both a publish date and an updated-at date, Google will go for the original publish date to display in search results. One publisher decided to test this and added both dates, which resulted in their blog CTR being knocked down 20%. After they removed the original date, Google was able to see the updated one and brough it back straight away. If you are regularly updating your older content, showing two dates might send mixed signals to Google and undermine the freshness you're trying to achieve.
Read

Google writes your headlines now
Google is testing AI-generated headline rewrites in Search, and publishers who have been watching what happened in Discover will find some similar patterns. Last year, what Google labelled as a "small UI experiment" turned quickly into a feature. Now Search is getting the same treatment, and Google is using identical language, that it is small, narrow, not approved for wider rollout. All things we've heard before. What makes this different from Google's existing title rewrites is that this new version will generate text that doesn't appear anywhere in the original article, so not pulling from your H1 or OG title, but Google writing its own version of your headline. No disclosure, no opt-out, and no way to see what's happening unless you check it manually. Of course, a headline that's good for Google might not be good for anyone else, and then we are off into the whole "So what is good for Google?" question.
Read

The March shuffle
Google has blessed us with a new Spam Update in March, the first spam update since August 2025 and the second algorithm update of the year overall. It was described as a standard update rolling out across all languages and locations, expected to wrap up within a few days. What it's specifically targeting isn't clear, but if your rankings or traffic have moved unexpectedly this week, this is a likely culprit.
Read

Stop collecting data, start using it
Most publishers are sitting on more audience data than they know what to do with, but that isn't the problem: the issue is that the data lives in multiple different places, and nobody has the full picture. Glide Nexa is built to fix that: a single audience interaction platform that pulls first-party data into unified profiles, connects editorial behaviour to commercial outcomes, and gives everyone a shared view of who their readers actually are and what keeps them coming back. Personalised recommendations, dynamic paywall messaging, habit-forming formats, and interaction and engagement signals - it's all there, without the enterprise price tag that puts such insight out of reach of most publishers.
Read

Big Tech

Europe to Google: wrap it up
A coalition of European publishers, tech companies, and industry bodies - including Axel Springer, News Corp, Condé Nast, and the European Publishers Council - have written jointly to the European Commission urging regulators to bring their antitrust investigation into Google's search practices to a close, and do it fast. The probe has been running since March 2025 under the Digital Markets Act, with a stated target of around 112 months to conclude. Obviously, the deadline has passed. The signatories argue that the delay is already doing damage, creating uncertainty which is affecting investments in both media and tech. On the other side, Google claims they have made changes to address the concerns. The Commission claims they plan to move swiftly.
Read

Fake Claude site, real malware
Imagine you're going through Google, searching for help with a Claude plugin, and after clicking on a sponsored ad at the top of the results and you land on a site that looks exactly like Anthropic's official documentation. The thing is, it only looks like the official thing and when following it the user will end up downloading credential-stealing malware onto their machine. The ad was verified by Google which means the advertiser passed the identity checks. Google did remove it after being tipped off, but it begs the question how many of these malicious ads are still out there.
Read

Who's Sora now
Disney signed a deal with OpenAI to invest $1 billion and license some of its characters for use in video-maker app Sora, with the goal of eventually integrating the technology into Disney+. That deal is now off, after OpenAI's decision to shut down the standalone Sora app. The entertainment company will still explore partnerships elsewhere, with a pointed note about finding platforms that "respect IP and the rights of creators", a jab that hits right into the launch of the Sora app.
Read

The case that couldn't stick
Two US news publishers - Helena World Chronicle LLC, and Emmerich Newspapers - lost their bid to sue Google for monopolising the online news market The publishers claimed that Google's grip on search traffic left them with no meaningful alternatives for reaching readers online. The judge rejected their argument as too flawed to hold up, as well as their inability to show that any losses qualified as antitrust harm rather than just the reality of operating in a Google-dominated market.
Read

AI & Copyright

Steal now, lawyer up later
A leaked video of a closed-door Stanford lecture has resurfaced, where the former Google CEO Eric Schmidt is telling students that if they need copyrighted content to build an AI product - they should just take it, and then hire lawyers to clean up the mess if the product succeeds. This is a blunt articulation of something that the AI industry usually dresses up in some fine legal language. The attitude he described isn't new, as the AI industry has trained most models on vast quantities of books and creative work without paying anyone a dime, while at the same time making sure that no one can train on their outputs. We can use yours, you can't use ours, the double standard runs deep.
Read

ISP ≠ Copyright police
The US Supreme Court has ruled unanimously in favour of Cox Communications, deciding that internet service providers can't be held liable for copyright infringement that their customers committed. Back in 2018, Sony and other major music labels sued Cox and argued that it should have cut off customers accused of pirating music. The court didn't agree with this, ruling that not disconnecting these customers doesn't make Cox liable for their crimes. Cox called this a huge victory, The Recording Industry Association of America (RIAA) doesn't really agree.
Read

Name dropping without asking
As our previous comment mentioned, Grammarly has quietly launched a feature called Expert Review, which offered writing suggestions framed as coming from real named people - S. King for your plot, C. Sagan for your science writing, and so on - but failed to mention that none of them agreed to this. Writers objected, a class-action lawsuit followed, and the feature was pulled after eight months. CEO Shishir Mehrotra sat down for an interview with tech writer Nilay Patel to apologise, and the whole episode leaves one question in the air: if a product slaps a real person's name onto their AI product, where is the line between inspiration and liability?
Read

Fair use? Not for everyone
Jack Conte, the founder of Patreon which helps creators get paid for their work, had a pointed observation about the AI industry's fair use argument: it fell apart the moment OpenAI signed deals with Disney, Condé Nast, Vox, and Warner Music. The argument was never that compensation is impossible, but that smaller creators have no leverage to ask for it. Conte isn't calling for AI to be stopped, but he does point out that the industry's selective generosity tells you everything about where the power sits.
Read

The AI Overview gap
AI Overviews appear in less than 6% of breaking news and major headline queries - a fraction of the rate seen in categories such as Health, Tech, or Business. Google seems to be cautious about generating AI summaries when news is moving fast, which does make sense considering AI hallucinations. What is more interesting is who actually gets cited:YouTube is at the top, followed by Wikipedia and Instagram, with traditional news publishers nowhere to be found in the top ten. NewzDash shares more info.
Read