DSA enforcement gaps, Ofcom's court win and another T&S vendor consolidation
Hello and welcome to Everything in Moderation's Week in Review.
A warm welcome to new subscribers from Cinder, Riot Games, Defend Digital Me, Ofcom, Trafalgar Strategy, TripAdvisor, X/Twitter, Runway and elsewhere.
It was brilliant to see so many of you last night at our Marked As Urgent event on the UK social media ban in central London. As one person said: "It was great to have an actual debate about the issue. And to start to imagine what a different framework looks like." More on that in Monday's T&S Insider and, in the meantime, follow us to be the first to know about future events.
If you ever have the urge to pretend you're a wrestler while on a panel, my advice is: don't.
Here's the big stories from the last seven days — BW
Checkstep’s final London Trust & Safety Meetup of 2026 is upon us!
Come along on Wednesday 2nd December 2026 for a festive gathering. This informal networking evening will bring together Trust & Safety professionals, policy leads, compliance experts, and industry partners.
Why attend?
- Meet and network with fellow Trust & Safety professionals
- Hear expert insights from industry speakers, including eBay
- Exchange knowledge and ideas in a relaxed, social setting
- Be part of building a stronger Trust & Safety community in London and beyond
Policies
New and emerging internet policy and online speech regulation
A new report from the European Commission's digital rules enforcer highlights the tension between enforcing the Digital Services Act and the ongoing efforts to set up enforcement across the bloc. Henna Virkkunen's annual update is understandably quick to point out preliminary findings against TikTok, Meta and Pornhub among others but also admits it is suing four countries in the European Court of Justice because their national regulators aren’t properly empowered to enforce the DSA. Almost three years in, that's got to be a worry.
Age old question: if you read Ofcom's rapid assessment report last week with interest (EiM #360), you'll want to read a new Commission analysis about the rights and business costs associated with age assurance. Interesting stat that I didn't know: the average age check will cost a platform £0.10 — with that number coming down from early regulatory estimates.
Important story to note from last week: Meta, X/Twitter and TikTok lost their case (EiM #358) against Ofcom regarding information requests they say go "beyond what is necessary or proportionate". The judge concluded that it was not "necessary to rely on any significant degree of deference to OFCOM’s own decision-making".
Also in this section...
- Major test for Australia's social media ban begins in High Court (BBC)
- Commission seeks feedback on advertising repositories under the Digital Services Act (European Commission)
- In Brazil, Lula's Army Fought Meta on Meta's Own Turf (Tech Policy Press)
- Where the Weight Lies: A Balancing Act for the Governance of Open-Weight AI Models (Lawfare Media)
Ctrl-Alt-Speech Patreon exclusive: Mike & Ben Review The Social Reckoning
In a special bonus episode for our Patreon supporters, Mike and I share our reactions to The Social Reckoning, Aaron Sorkin’s much anticipated (but it seems not much loved) followup to his acclaimed 2010 movie The Social Network.
I wrote about how the film acts like $30 million of media literacy but safe to say, Mike doesn't agree. Whether you've seen it or not, you'll enjoy this one.
Products
Features, functionality and technology shaping online speech
Compliance managers and lawyers, take note. A new product from age assurance technology provider k-ID claims to help builders "never think about regulation again". Announced last week, Kai integrates product compliance as you code and claims to suggest changes when new regulation comes online. The impressive launch video shows how the tool defines the age gating for a nine-year-old in Brazil without knowing anything about the Digital Child and Adolescent Statute.
I've said before that we're going to see T&S vendors become consolidated as the market responds to regulation. This week we saw another example of this as Checkstep, an AI content moderation platform, acquired Unitary's T&S business for an undisclosed amount. Unitary — which, like Checkstep, is based in London — previously raised $15m in Series A funding back in 2023. (Full disclosure: Checkstep happens to be the sponsors of today's newsletter but has no input into the content).
Also in this section..
- Agile Deliberation: How AI helps humans define the gray areas (Google)
- Apple's new 18+ app limits mean little for dating apps, a lot for AI companions (Mashable)
- The Incident Report Is Becoming a Building Block of AI Governance (Tech Policy Press)
- Covert underperformance on safety research tasks (OpenAI Alignment)
Platforms
Social networks and the application of content guidelines
It was inevitable that greater attention would be paid to OpenAI's safety processes and teams in the light of the Tumbler Ridge shooting (EiM #326) and we've seen that in various guises this week:
- An investigation from the Observer looks into the way the company set up its safety team and speaks to an anonymous member of its intelligence and investigations team who has been "frustrated by her team's lack of resources".
- Josh Lawson, the company's head of civic and strategic risk, told Politico that its "Elections Operations Center room" would help "respond quickly to emerging issues and enforce our policies as people head to the polls" — although I'm not sure I'd house it in a "small conference room off the cafeteria".
- The company itself has been pumping out news releases on everything from its new protections for teens to its partnering with an external firm to improve its security — suggesting an attempt to better explain its safety work.
Apple announced that it will adjust its App Store policies to make apps that enable "relationship connections between real people and/or a virtual companion that engages a user in sexual or romantic content" only accessible to users over 18. I would've thought that was already the case for both categories so this feels like the company catching up with user expectation — not to mention regulators wanting Apple to play a stronger role in child safety (EiM #360).
Also in this section...
- Meta Streamlines State Censorship in India with Automated Takedowns (Tech Policy Press)
- Meta Is Refusing to Remove AI-Generated Videos of Coaches Inappropriately Touching Young Athletes (Futurism)
- Instagram Sex Trafficking Lawsuit Partially Revived on Appeal (Bloomberg Law)
People
Those impacting the future of online safety and moderation
It's been a while since Prince Harry and Meghan Markle cropped up in EiM. The last — EiM's People Directory tells me — was back in August 2024, when the couple launched "The Parents' Network" for people who have suffered in some way from online harm — whether it be cyberbullying, sextortion or suicide. The network was handed over to a US non-profit earlier this week, for reasons that are unclear. The former working royals have made several child safety speeches in the last 18 months too.
Now, the pair are coming to Brussels. Politico reported they will take part in an exhibition organised by their foundation to commemorate "the memory of children whose lives were lost in connection with online harms" at the European Parliament. The Lost Screen Project contains a host of tragic stories, narrated by their parents, some of which have very direct connections to social media and others that are more tangential. How Harry and Meghan reflect the nuances of these cases next week will be interesting to see.
Check out the full list of people in one place, with a link back to the edition they appeared in > EiM People Directory
Posts of note (jobs edition)
Handpicked posts that caught my eye this week
- "Holy moly, $500K. For a fraud role!" - Brian Davis finds that a) Polymarket might actually be taking fraud seriously and b) old harms are the new harms.
- "As a Project Manager, you’ll play a key role in taking forward our mission" - The Centre for Information Resilience is looking for someone to join its impactful team.
- "So when OpenAI engaged me and my partner Nicolette Hartman, we asked the question we always ask: who do you want? Their answer: "The one person who wasn't convinced to join us."" - a fascinating insight into how the frontier labs think about reputation and talent from Paul Daversa.
Member discussion