4 min read

📌Platforms release new safety features, Brussels gives go-ahead and moderating Arabic

The week in content moderation - edition #138

Hello and welcome to Everything in Moderation, your regular digest of the most important reads on content moderation, published every Friday. It's selected and scribed by me, Ben Whitelaw.

A hearty hello to new subscribers from Spotify, Collate, IMG, Simon Fraser University, Niantic Labs and a host of people from Cornell University. I’ve connected with some brilliant people through EiM so, whether this is your first edition or your 100th, feel free to drop me a line to say hi and explain how you got here.

A nugget from last week's newsletter that is perhaps very telling: I used 'metaverse' in the subject line and the open rate dropped a full ten percentage points. I'm taking it to mean that you're interested in the problems we face here and now, rather than the ones we're yet to encounter.

Today's links are about exactly that (particularly my read of the week). Here they are — BW

📜 Policies - emerging speech regulation and legislation

The big news this week came from Brussels as European ministers gave the go-ahead on the Digital Services Act, which will govern harmful content online, and the Digital Markets Act, which addresses aspects of platform competition. The agreement at the quarterly Competitiveness Council makes it likely that the DSA's final text with be signed off early next year, less than 18 months after it was first presented by the Commission. As Politico reporter Laura Kayali notes, that's lightning quick.

It wasn't all plain sailing though. A plan put forward earlier this week to provide media outlets with an exemption was pushed back on, according to Euractiv, with European Commission vice-president Věra Jourová going as far to say such measures would be one of “good intentions leading to hell.” I'm particularly interested in this dance between regulation and press freedom, having written a piece for the British Journal Review back in September, building on the work of others, showing that the Online Safety Bill's own media exemption causes as many problems as it fixes. My read of the week.

Saudi Arabia is moving forward with its attempts to regulate digital platforms and published a consultation on its draft bill earlier this month. If you're a close watcher of regulation efforts elsewhere, there's a lot that is familiar: as Access Partnership's Hussein Abul-Enein points out, the Kingdom seeks clear processes for takedowns and expects platforms to set up an office in the country if it doesn't have one. More than other regulatory efforts though, this bill is likely to cause headaches for platforms because of the poor record of moderating Arabic and censoring dissenting activists across the Middle East (EiM #136). For organisations planning to submit to the consultation, act quickly - the deadline is this Tuesday.

Notes or no notes? That is the question in Ireland, where Minister of State Robert Troy has been meeting with tech companies in lieu of the Digital Services Act. The Business Post reported this week that Apple, Google and Facebook asked for "no detailed notes to be taken", which would have been subject to Freedom of Information requests, whereas TikTok made no such request leading to reports of its discussion with minister Troy.

💡 Products - the features and functionality shaping speech

I missed this when it was announced earlier this month but Spotify has made it easier to block users who stalk or harass. The feature, which has been "in consideration" since 2018 and can be found via a user's profile, prevents selected users from accessing your listening activity or public playlists. As many have noted, it's about time.

Five projects have been awarded £85,000 as part of the UK government's efforts to find solutions to the challenge of finding child sex abuse material (CSAM) while maintaining end-to-end encryption. Age assurance provider Yoti and Crisp Thinking, a risk intelligence company, were among the successful projects for the Safety Tech Challenge Fund and now enter a five-month deliverability phase. Which, coincidentally, is around the exact time that the Online Safety Bill will go in front of MPs in the UK's House of Commons. Nadine Dorries will be hoping they can come up with something.

While we're on the topic of government funding for content moderation start-ups, CaliberAI — a warning system for harmful and defamatory content that received Enterprise Ireland funding to get started — has signed a deal with Mediahuis Ireland, which publishes Independent.ie and the Belfast Telegraph. Mariana Mazzucato, whose brilliant work on "public value" is worth digging into, would be proud.  

💬 Platforms - efforts to enforce company guidelines

Five months after it was banned for "immoral/indecent content", TikTok has finally been restored in Pakistan. The Pakistan Telecommunication Authority (PTA), which oversees online platforms, released a statement last Friday confirming that the Bytedance-owned video app had agreed to "various initiatives being taken along with a future strategy and investment in the local market". As of April 2021, TikTok had been downloaded 39m times in Pakistan.

Are you a subscriber to Twitter Blue, Jack Dorsey's new subscription offering for his 280-character chronicler of chaos? I'm not (although I would if it introduced this one feature) and I'm guessing you're not either. Bloomberg's Parmy Olson writes that she would "happily pay Twitter if it would do one thing: hire more people to weed out spammy content and harassment".

In not-unrelated news, Twitter's Safety Mode‚ announced in September and autoblocking for seven days, has started being rolled out to users.

👥 People - folks changing the future of moderation

In a week that yielded significant regulatory news from Brussels, it seems right to appreciate the work of those reporting from the corridors of the European Union. One media outlet, in particular, deserves mention.

Politico Europe was only founded in 2015, some eight years after its US counterpart, but quickly delivered on a promise to "prove that news about the European Union doesn’t need to taste like oatmeal."

I've begun to read it more over the last two years as plans for the Digital Services Act emerged and that's largely down to its team of tech and policy journalists that bring a steady supply of scoops (see the recent piece on UK minister Caroline Dinenage from EiM #136). There are plenty of people I could mention but I gravitate toward anything written by Samuel Stolton, Clothilde Goujard and the aforementioned Laura Kayali. All are worth following if you don't already.

🐦 Tweets of note

  • "FB is divesting from human moderators, content moderation AI does not work, & we depend on emailing FB's human rights team on a Sunday -knowing that they will only see it on Monday." - take your pick from this list by Berhan Taye on why the current system is not working.
  • "Translation: "I want to be able to wield power with impunity" - tech lawyer Cathy Gellis rightly throws some shade at UK elected representative, David Davis.
  • "We are looking for line managers/people who feed into trust and safety policy for CSAM content moderators to take part in an online workshop" - interesting opportunity from the Centre of Abuse and Trauma Studies at Middlesex University.