(AI illustration)
October 28, 2026
There are times when you think that Meta’s Mark Zuckerberg has no interest in shedding his cred as Big Tech’s baddest boy.
Last week Brazil’s Lula government accused Meta of interfering in the Presidential election by pumping the algorithmic tires of his right-wing rival Flávio Bolsanaro while suppressing incumbent Luis Lula da Silva’s social media content.
Zuckerberg denies the allegation.
Unfortunately his reputation precedes him: no one has forgotten Meta’s political interference in the Cambridge Analytica scandal that cost Meta billions in fines. Or Meta’s years of hosting and promoting toxic content on Instagram and Facebook. Or, closer to home for Canadians, the Meta sabotage of the Canadian Online News Act with its content ban on mainstream news organizations.
Zuckerberg’s latest caper is an update of its take-it-or-leave-it terms and conditions for Facebook and Insta business accounts, including the accounts of news organizations.
One significant change is Meta explicitly claiming the right to ingest all media content for the purpose of AI training. The word “training” was removed after protests, but it remains implied in the grammar of the new legalese. A bigger change was the widened scope of the content grabbing: not just posted articles but allegedly the entire website associated with the account holder.
News organizations would then have a choice: get off of Facebook and Instagram or surrender all of your online content, including the archive.
A caveat is that Meta has made a non-binding promise in a press release to respect crawler-blocking robot.txt instructions used by website-hosts to block scraping. But technically that scraping block can be evaded by crawlers and Zuckerberg’s promise is not part of the revised terms and conditions, probably because it would neutralize the content grab that Meta is reaching for.
It’s a global play by Meta and the hornet’s nest has been well poked. As the November 20th deadline for the escalated crawling approaches, news organizations have a decision to make.
A National Post opinion editorial co-written by News Media Canada’s Paul Deegan confirms most Canadian online news outlets are affected. But of course mainstream outlets are already blocked from posting news content on their own accounts, so their account pages look like this:
***
In MediaPolicy’s recent comment on the Safe Social Media Act C-34, we mooted the wisdom of front-loading more of the expected child safety features in the legislation instead of delegating the task to a digital regulator that will take months to set up, more months to consider safety codes that social media companies will likely under-design anyway, and yet more months for the regulator to review and impose stronger measures.
And in the meantime the bill requires a temporary ban on children’s accounts that a majority of kids will evade anyway.
In a Substack column, Andrew MacDougall expressed a public policy impatience that would probably resonate with most parents:
A different debate would be had if the government were to come out and specify some product design features they don’t want to see in children’s social media accounts. After all, most of the harms follow on from the product design features of social media platforms. The reason kids have been shown self-harm content is because kids express an interest in that content and the algorithm then serves them more. The reason some kids obsess over their social media output is because the platforms have social validators in the form of like and share counts.
But what if we binned off those product design features? What if the government told the platforms removing those features was a precondition for being considered safe? You can’t sell a car without seatbelts or a drug that doesn’t pass trials, so why can you flog a platform without comparable safety features?
Put differently, would we need suggested daily time limits if the social media platforms didn’t have product design features like infinite scroll, auto-play for videos, or algorithmic recommendation systems? What if they gave us something closer to OG social media, when all we had was a reverse chronological feed of content from the people we chose to connect to? Would that be safer? It would certainly be more boring and less lucrative for the platforms.
In other words, why not write some tough standards for social media companies right into the bill?
Any overreach could be cropped back later and, as incentive for social media platforms to make their first pitch their best, an interim safety code might be preferable to an interim ban on children’s accounts.
***
When I last wrote about AI in newsrooms I meandered through some potential innovations in audience tools and platform monetization we might see in the near future, assuming rogue AI agents haven’t scorched the Earth by then.
At the UK’s Daily Telegraph, the new German owners Axel Springer have pulled back the curtain on big changes including newsroom layoffs, improved efficiency something-something, and —of greater interest— the “reinvention of the reader experience” by relying on technology, data and AI.
There wasn’t much more detail provided by Telegraph management than a directional statement. There was a cryptic reference to focussing on new revenue streams including “video, audio, puzzles, events, B2B, licensing, commerce and new products.”
If that statement is intended to convey an AI-based plan to reconstitute “the bundle” of news and non-news content at which newspapers excelled prior to the content-fragmentation of the Internet, the fate of the The Telegraph will be something to keep tabs on.
As I mentioned in my previous column, Canada’s own Village Media is setting an example in developing its own AI interface to connect readers to community resources.
This raises the question of whether AI can be more than just a threat to wipe out news organizations with Chatbots and more than just a tempting short-cut for newsrooms and journalists.
The larger news outlets have the capital and human resources to succeed at this. Just to take another example, the Wall Street Journal used an AI tool to ingest 25 million TikTok videos and 500,000 podcasts to analyze news content reaching voters in the US mid-terms.
You could see the Globe and Mail or La Presse imitating or even escalating this kind of innovation. Historically they have been the apex innovators in Canadian news media, along with Village Media. But what about the metropolitan or local news outlets who hardly have the time, money or expertise to get on the AI innovation train? It’s hard to sprint in the middle of a marathon.
To segue from this point, I recently reviewed the federal Heritage Department’s report on the consultation it did with Canadian publishers on the administration of the Canadian Periodical Fund and the Local Journalism Initiative, two programs that support local news with newsroom subsidies.
What I discovered in its 28 pages was the industry participants’ unfortunate preoccupation with reapportioning subsidies from large and group publishers to the benefit of small and independent outlets.
What was missing was a preoccupation with innovation, although it did not go completely unmentioned.
The lack of enthusiasm was perhaps not surprising given that Heritage did not renew its short lived Business Innovation Fund in 2024 because —I was informed by an industry source — of undersubscription. In the new report, the idea of an AI Skills and Experimentation Fund was raised by magazine publishers, not news outlets.
As AI rolls over the news media industry, it really looks like Heritage should be considering how to play matchmaker between local news organizations and AI tools and products.
***
If you would like regular notifications of future posts from MediaPolicy.ca you can follow this site by signing up under the Follow button in the bottom right corner of the home page;
or sign up for a free subscription to MediaPolicy.ca on Substack;
or follow @howardalaw on X or Howard Law on LinkedIn.
I can be reached by e-mail at howard.law@bell.net.
This post is copyrighted by Howard Law, all rights reserved. 2026.
