Ink and income, Google got played, infinite scroll on trial, and AI gets a backstage pass

From tweet to publisher, the encyclopaedia fights back, Meta unlocks the DM door, and Altman speaks, sort of - all in this week's Content Aware.

Published: 15:12, 19 March 2026

Corbidge comments on… a cautious win for creators
Government enthusiasm for a new way of treating copyrighted work and its use by AI, which would have driven a coach and horses through existing norms and the rights of creators, seems cooled - for now. A plan to require rights holders to formally request any and every AI bot to stop stealing stuff has been ditched, but as our journalism jockey Rob reports, the war is far from won.
Read

Publishing & Media

Record journalists, record revenue
While some newsrooms are cutting staff, The New York Times has adopted the opposite strategy and reached a record headcount - a whopping 2,300 journos and growing, 50% more than a decade ago. According to its most recent annual report from 2024, the growth is backed by a well-segmented revenue model well beyond subscriptions, including games, cooking, sports reporting, and product recommendations. We're talking nearly 13m subscribers, $802m in quarterly revenue, and a target of 15m subscribers by 2027. The contrast with recent gloom in some sectors of the industry couldn't be starker, as the NYT shows that quality journalism is still a viable business model.
Read

90 day AI
Imagine this: 20 brand new websites, 2,000 AI-generated articles, zero human editing. Within 36 days, 71% of pages were indexed, 8 sites ranked for over 1,000 keywords each, and the content was pulling real traffic. No domain authority, no E-E-A-T, none of the signals Google usually claims to care about. Then the trouble started and every article vanished from the top 100 as if it was never there. This goes to prove that ranking without a real SEO foundation is just borrowing time, and that tools which pump out content, and good content, are two very different things.
Read

Jokes on Google
Irony alert after reports of a Google core update did the rounds amongst SEO circles, before being revealed as a hallucination spread by - Google AI Overviews! Jon Goodey, an SEO expert, decided to test what would happen if he were to trust a hallucination and publish it as if it was real to his feed - like many people do. Google ranked it, AI Overviews surfaced it, and other SEO sites picked it up and ran with it without batting an eye. This only confirmed what most people already suspected, that while there is no meaningful fact-checking on search results and it leans heavily on the reputation placed in certain key algorithm-driven sources, it only takes one of those systems to wave a fiction through for the others to assume it's likely true too.
Read

The X factor
Google Discover has recently shown a lot of love to X and surfacing cards in results, but in a peculiar break from normal it's site owners getting the traffic not the ex-Twitter. If the post is a link post type, rather than a standard post with a URL pasted in, the Discover card links directly to the publisher's site, cutting out X entirely. This turns X into more of a distribution layer that feeds clicks from Google Discover to publishers. Have a look at how you mark up your posts to see if you can benefit.
Read

GEO can't save a broken SEO
The GEO hype is running high, and even if it sounds helpful, a lot of the advice around it doesn't really stand up to scrutiny. We're talking rapid AI search skipping the part where rankings crash shortly after, losing search traffic and potentially getting pulled out of AI responses. SEO expert Lily Ray shares more on whether AI search is built on the same foundations as organic search, and if GEO is just SEO with a new hat on.
Read

Make readers stick around
Ever thought about what UK's most engaging news apps have in common? They wrap a lot more than headlines around their readers, tying in puzzles, audio, video, infographics, tools, and much more. The apps pulling the highest hours per user per month are the ones that give people reasons to stay beyond the news cycle. Glide's very own stick in the mud Rich Fairbairn breaks down what separates the products that keep readers coming back from the ones that don't, drawing on habits of the teams consistently topping the engagement charts.
Read

Big Tech

Britannica joins the sue cruise
Encyclopaedia Britannica his suing OpenAI, alleging it used nearly 100,000 of its written articles to train ChatGPT without permission so it could reproduce near-verbatim copies of its content in responses and hoarding traffic. The lawsuit also accuses OpenAI of trademark infringement, alleging it implied to users it had permission to use the material, as well as citing Britannica as a source in AI hallucinations just to rub salt in. OpenAI's standard lawyer-approved response circled around publicly available data, fair use, innovation, and so on. This case joins a growing pile of similar lawsuits from authors, news outlets, and reference publishers, all asking the same questions that courts have yet to answer.
Read

Swipe, repeat, repeat
A landmark trial has just wrapped up its closing arguments in Los Angeles, where two of the world's biggest tech platforms were accused of deliberately engineering their products to hook children. Infinite scroll, autoplay video, and notification alerts were all put under the microscope, and a former Meta employee describing them as promising something better just a swipe away, while there is no upper limit built in. Both Google and Meta are saying the features are just giving users what they want, while the verdict to this trial could reshape how courts think about platform design and the responsibilities that come with it.
Read

Hallucinations of Sam
The Onion sat down with Sam Altman for a chat around ethics, military contracts, job displacement, and the purpose of humanity. Or did it?
Read

Private DMs? Not so much
Starting May 8th, Meta is killing end-to-end encryption on Instagram, burying the announcement on a support page with no fanfare and no press release. The official explanation is low adoption, but, yeah, it's money - either the lure of your conversations training AI, or the threat of fines or some other reverse-inducement from someone bigger. If you go back to 2019, you'll see Zuckerberg spend over 3,000 words making the case for encryption being a fundamental privacy right, naming it the future of the internet.
Read

AI & Copyright

Seedance in dubious copyright waters 
ByteDance's new video model Seedance 2.0 went viral for the wrong reasons. Before it even had a global release, users started generating fake celebrity fights, complete with shaky physics and weird cameos, and as you can expect, movie studios and streaming services were unimpressed. While ByteDance says it's still working through content restrictions and copyright compliance, anonymous sources claim the legal action from rights holders is what's behind the freeze on an international rollout. This is the pattern seemingly: a model shipped on a legally dubious use case, users use it in predictable fashions, there is a swift lawsuit, and a renewed copyright debate as the cherry on top.
Read

AI Overviews eat a bunch of clicks
Google's AI Overviews now appear on over 43% of UK searches and 38% in the US, contributing to a drop in organic results CTR dropping from 27% to 11%, according to SISTRIX. Being cited in AI Overviews doesn't really help too much, since most users don't bother clicking through. According to our friendly SEO expert Barry Adams, in order to come out on top, publishers need to offer the readers something that AI can't. For now, publishers still have breaking news, as AI can only react to those rather than report it, but it is obvious that with the way AI is developing, more firepower will be needed.
Read

Hollywood's AI pivot
How the tables have turned at Tinseltown, as now Hollywood has decided it won't fight AI anymore, but build it into production pipelines. Major platforms such as Netflix, Peacock, and Prime Video are all actively integrating it across workflows, for tasks such as mapping scenes before filming and generating near-final visuals, framing the whole thing around a human being in the loop at every stage, partly because copyright protection requires it. Cost savings are being redirected into creating more content and not smaller crews, for now at least, as SAG-AFTRA negotiations are still going on. 
Read