The quotes nobody said, Europe's digital bouncer limbers up, and back off the back button

The one-man misinfo hunter, creator journalism's money problem, opt-out ignorance, and the art heist nobody stopped - all in this week's Content Aware.

Published: 15:15, 16 April 2026

Corbidge comments on…same water, different rules
Both the New York Times and The Economist have been doubling down on what makes then distinct, with the former tightly controlling what content escapes the paywall, while the latter reinforced their anonymous "hive mind" authority in an era where personality-driven media is king. The lesson for smaller publishers? Platforms are far from your friends, self-dealing trust signals do matter, and the publications that rise up and survive did so because they refused to bow to convention. Our content crusader Rob explores what the rest of us can learn from these two titles that went for defiance over compliance, and why your gift articles should never go to waste.
Read

Publishing & Media

The quotes no-one said
If you are going to use made up quotes and facts, doing it in a newsletter about the media, sent to the media, quoting people from within the media, is probably not the best place to do it. Mediaite's paid-for aggregator email One Sheet has been accused of letting AI fabricate quotes, misattribute facts, and mix up people and organisations, with trusted news sources including Status, Press Gazette, and NiemanLab misused. The author blames dodgy data from a flaky spreadsheet rather than AI hallucinations, but has been suspended while under investigation. Whether AI-generated or human error, the result stays the same: a newsletter intended to reinforce trust in journalism instead diluted it.
Read

The illustration nobody asked for
The New Yorker commissioned an AI-generated illustration for its profile of Sam Altman, and - perhaps unsurprisingly, given the publication's usual stance - prompted a vigorous discussion around the whole use of AI in art and illustrations. The commissioned artist is renowned for using AI amongst other tools, going so far as to code his own, and would it not be a knowing and ironic take for a magazine to use AI to illustrate a story about AI? But critics say it still looks like AI, and using the aesthetic of AI slop to illustrate a piece about a man responsible for AI slop is too close to the bone to be clever. For illustrators seeing art budgets slashed, it's less about the art and more about the normalisation of bad.
Read

The hustle no-one tells you about
A survey of 43 independent US news creators found that only five can make a living from content, while most make next to nothing. The report, commissioned by the Center for News, Technology & Innovation, paints a picture of mission-driven journalists, who either left legacy newsrooms or started out alone, and who work long hours alone relying on subscriptions, freelancing, consulting, savings, and partner support to get by. Basically, the life of a freelancer, countless predecessors would say. Other insight points to stress, imposter syndrome, reluctance to ask for money from audiences, and fundamental beliefs in information being free. As one creator has put it: "Writing for a wealthy group of people is the only way at this point to run a media business." NiemanLab tells more.
Read

The misinformation audit trail 
Walker Bragman runs the Accountability Journalism Institute, a small non-profit dedicated to tracing where misinformation begins, who pays for it creation or distribution, and why. His focus is mostly on public health, having previously investigated the likes of Fidelity Charitable and the Brownstone Institute during probes into misinformation during the COVID era. The operation runs on small donations and reader contributions, with most of the reporting by Bragman himself. Editor and Publisher shares more details.
Read

New owner, new backdoor
Dozens of WordPress plugins have been pulled offline after the discovery of a backdoor in their code. The Essential Plugin affected over 20,000 active installs into sites, and the "attack" route was somewhat unusual: Essential Plugin was acquired by a new owner last year, a backdoor seemingly was added to the source code which lay unactivated until earlier this month when it started pushing malicious code to website users. As WordPress can't notify users about the change of ownership, the attack went undetected. It's the second WordPress plugin hijack in two weeks, so if you run sites on WordPress do check the list of affected plugins.
Read

The hateful backjack
If you've ever clicked the back button and ended up anywhere but on the previous page, then you've experienced back button hijacking. Google has now marked it as a violation of its malicious practices policy and gives site until June 15, 2026 to remove the technique or face manual spam actions and automated demotions in search. The annoying bit for some site owners is that they themselves are not aware of it as the "feature" is introduced by advertisers. Additionally, it's not clear what will happen if the origin page is a feed that loads fresh each time, which to many is just as annoying when they are presented with effectively a new beginning.
Read

Prove your audience is paying attention
As engaged traffic and passing traffic becomes harder to differentiate, advertisers and site owners want proof audiences exist and are genuinely engaged. If the engagement proof points don't exist, or are logged in multiple disconnected, it can make it impossible to quantify and segment them for advertisers. That's where Glide Nexa comes in, tying every interaction to a known user profile in real time, building first-party engagement data that publishers own and can present to advertisers with the data they need.
Read

Big Tech

Taking the Mythos
When Anthropic restricted the release of its mythically scary new Mythos model to a handpicked group of tech companies and US government agencies, plenty felty left out - like national and international regulators. Politico contacted eight European cyber agencies, and only Germany's could say it had spoken with Anthropic, while none have been able to test the model - so far that's limited to a dozen mates in California or on the tech biz circuit. The UK's AI Security Institute had already tested it and published its assessment, while the EU AI Office declined to say whether Mythos had even come up on the schedule.
Read

Prove your age, keep your data
An "EU-ID" age verification system for digital life is "technically ready", the bloc says, with France, Spain, and Italy among seven member states who plan to integrate it into digital national ID wallets. The app uses zero-knowledge proof, so users can verify age without having to hand personal data to sites or social platforms. EU President Von Der Leyen's message to platforms was "no more excuses". The system code can be independently verified, with the goal of an EU-wide system rather than a patchwork of national ones. No timeline for full deployment yet and no regs on whether platforms will be required to adopt it, but with numerous territories looking into social media bans for young people, the direction is clear.
Read

The ban that banned nothing
After four months of a social-media ban for under-16s in Australia, the results aren't groundshaking, for good or bad. According to a study by the Molly Rose Foundation, 61% of Australian 12-to 15-year-olds with accounts prior to the ban still have those accounts. The platforms don't seem to have done much to enforce or encourage change, with 64% of YouTube users, 61% of Snapchat users, and 60% of Instagram and TikTok users having taken no action to remove their accounts, and workarounds are easier than most people for or against predicted. Half of those surveyed believe that the ban made no difference to their online safety, while 14% say that it made them feel less safe.
Read

Privacy settings: ignored
Yes, they ignore your choices. An independent privacy audit of more than 7,000 popular websites found that 55% of sites set ad cookies in a user's browser even after they opted out of tracking. Google was worst, ignoring opt-out prefs 87% of the time, followed by Meta at 69% and Microsoft at 50%. Yikes. The violations are hiding in plain sight in network traffic: the browser sends out the opt-out code, but the server responds by setting an advertising cookie anyway. Google's own certified consent management platforms failed the test as well, with opt-out failure rates going from 77% to 91%. All three companies dispute the findings or blame misunderstandings in how their system works. Weird how those misunderstandings are never to the detriment of the companies.
Read

xAI vs Colorado
Elon Musk's xAI has filed a lawsuit against the state of Colorado for blocking a 2024 law that requires developers of "high risk" AI systems to prevent algorithmic discrimination in areas such as housing, education, and employment. xAI claims this puts excessive burdens on developers, and its addled chatbot Grok which might have to be neutral in such fields instead of giving jaundiced answers. In the AI state or federal regulation debate, if xAI wins, it could discourage any similar laws. If they lose, it could become a template.
Read

AI & Copyright

Using more, trust less
In a recent report Goldman Sachs found no meaningful relationship between productivity and AI adoption at the economy-wide level, even as 70% of S&P 500 management teams discussed it on earnings calls. 1% quantified its impact on earnings, while a Boston Consulting Group study found that 14% of enthusiastic AI adopters are reporting "AI brain fry" - mental fog, difficulty in focusing, and slower decisions - which hit entry-level workers the hardest. A third of consumer believe they don't need AI at all, while a New Gallup poll tells a similar story: roughly 3 in 10 workers use AI frequently, most report productivity gains, but half use it once a year or not at all. Of those that have it but don't use it, 46% simply keep working the way they already do.
Read

Invisible visitor eating your content
The scraping infrastructure behind bot visits, which went from 1 in 200 to 1 in 31 in 2025, is more sophisticated than most thing. Residual IP proxies, headless browsers, spoofed referrers, and services that scrape Google's search results are all at fault for publisher content being on a platter, without the bots ever needing to access the publishers' servers. Testing across 30 high-authority publisher sites found that some scrapers grabbed the full version of paywalled articles, not one checked robots.txt, and all went through as regular Chrome traffic. Google spent millions for a tool to stop one scraping company, and they still got around it. For publishers that don't have the resources Google has, the math isn't looking too good: the volume of scrapping is getting higher and higher while the return to publishers is zero.
Read

Billions of images, zero consent
In 2022, AI image generators scraped billions of images from the internet without credit, compensation, or consent. Fast forward three years, entry-level illustration jobs are no longer viable, creative workers are out of work, and the culprits' responses have ranged from indifference to contempt. The fact that OpenAI's former CTO suggested that jobs destroyed by AI "shouldn't even been there in the first place" didn't help the case either, but it showcases what Big AI thinks about your employment prospects or what you might be seeing and reading.
Read