Fresh insights for content moderation, combating disinformation, fact-checking, and promoting free expression online
Moderating the Marketplace of Ideas
Sadly this month’s news brings us more hate. Hate from online speech that leads to real-life violence against Asians to celebrities forced off social media because of harassment and even to a U.S. Supreme Court justice arguably advocating for more toxic content. Technology companies are trying to address the issue, but their ideas are not always well-received. It all makes for some interesting reading.
🧑💻 How Anti-Asian Activity Online Set the Stage for Real-World Violence
Increase in offline Anti-Asian hate crimes is linked to the online anti-Asian rhetoric content, particularly on messaging apps and online forums such as Telegram and 4chan.
✍️ Users can appeal Facebook's editorial decisions
Facebook’s Oversight Board will now start accepting user appeals if the platform fails to satisfy users with respect to content decisions. Previously the platform only accepted the Oversight Board’s decision on content restoration but now the Board will also have jurisdiction over content takedowns.
😶🌫️ For Political Cartoonists, the Irony Was That Facebook Didn’t Recognize Irony
Facebook seems to be missing a funny bone. A satirical cartoon aimed at the Proud Boys by left-leaning cartoonist Matt Bors was taken down by Facebook for violating its policies. Facebook’s policies currently allow parodies and satire, but such content is often taken down as their artificial intelligence systems and human moderators fail to understand the intent of the content.
⚖️ Analysis | The Technology 202: Clarence Thomas takes on social media companies' power
The Supreme Court’s staunchest originalist favors requiring tech companies to host hate speech. Justice Thomas, as he often does, stood alone on this one, though.
🎮 Intel's new Bleep software uses AI to censor hate speech
Dial up (or down) hate speech? Intel recently developed new software, Bleep, in partnership with Spirit AI to help tackle hate speech on gaming platforms. However, instead of completely eradicating such content, the software gives users the option to choose the level of hate speech they want to be exposed to.
📈 Google-owned YouTube reveals a new content moderation statistic
YouTube has added a new metric to its transparency reporting. The “violative view rate” is an indicator of the number of times people have watched content that steps over the line in violation of YouTube’s rules. Seems a little defensive, but, hey, they do get a lot of flak.
📱🙅♀️ Twitter’s "unofficial mayor" Chrissy Teigen quits platform after years of harassment
After facing online abuse and targeted harassment for years, Chrissy Teigen, a popular Twitter user, has left the platform. If platforms don’t effectively moderate content, could this be the start of a public movement?
💻 One in three spot hate speech in online videos
A recent Ofcom report highlighted that at least 32% of users of online platforms have been subjected to hate speech. To support this, Ofcom released guidance for video-sharing platforms urging them to increase their efforts towards targeting online harms.
Making Presidents blush. Not only Big Tech, publication platforms such as Medium also have faulty recommendation algorithms. The platform’s algorithm kept recommending erotica pieces to President Biden.
👀 Trump's Facebook Ban Will Likely Be Overturned by New Oversight Board
The Oversight Board’s recent rulings have made people question whether Facebook’s new Oversight Board actually has online safety in mind. Some observers have even called it a “clever sham”.
⚽ It was time to make a stand - Henry
Following the recent incidents of social media abuse of sportspeople, former Arsenal and France striker Thierry Henry quit social media in an attempt to end social media racism. This movement is also supported by current football stars such as Gareth Bale.
🎧 Clubhouse defined a format — now it has to defend it
Ever since the launch of Clubhouse, several platforms, including Big Tech, have tried to mimic the idea of having an audio-based social network. Clubhouse faces competition from platforms like Twitter and Facebook, who are “experts” in content moderation. With recent incidents of data leaks, Clubhouse is under pressure to keep up with the industry’s best practices in terms of effective content moderation.
🔞 Pornhub Just Released Its First Transparency Report
The website released its first transparency report after enforcing its new trust and safety rules on the platform and banning unverified uploads and downloads. The site removed 653,465 pieces of content that violated its Terms of Service in 2020, according to the report. The report is first of its kind for any platform hosting adult content.
Remedying COVID-19 and Vaccine Misinformation
🗣 Far-Right Extremists Move From ‘Stop the Steal’ to Stop the Vaccine
Far-Right extremists have moved to platforms like Telegram to propagandize the adverse effects of COVID vaccines claiming that through vaccination drives, the American government aims to “exercise excessive control” over the public. Essentially, shifting the narrative from #StoptheSteal to #StoptheVaccine.
💉 ❌ Facebook has an apparent double standard over COVID-19 misinformation in Brazil, researchers say
The Venezuelan President's Facebook account was recently frozen for spreading COVID misinformation, but the Brazilian President's account still remains live and actively makes similar posts. Facebook’s lack of action has people wondering if its content policies are just not applicable to Brazilians.
📈 🖥️ The vaccine misinformation battle raging in France
A BBC Monitoring report recently highlighted the rise in the spread of COVID misinformation in France. A Facebook page known to post such content saw an increase in the number of likes from 3.2 million to 4.1 million.
Regulatory News and Updates
🇨🇦 Canada is Bringing in New Legislation to Stop the Spread of Online Hate. Here's How It Can Work.
Canada finally aims to introduce regulations to address online harm, after having no such regulations in the place before. However, human rights experts see it as a way of limiting freedom of expression.
France recently implemented a “pre-transcription version” of the Digital Services Act, before the implementation of the actual regulation, thereby angering all the other EU member countries.
🇹🇷 Twitter will set up a legal entity in Turkey to comply with controversial social media law
Twitter appointed a legal entity in Turkey, to comply with the recently introduced Internet Law, where in addition to hefty fines for regulation breaches, major social media networks would need to appoint a local representative to address content takedown requests. This decision came after the platform was charged a fine of $5.1 million last year.
🧑⚖️ Google, Facebook Twitter grilled in US on fake news
During the House hearing for the on-going spread of disinformation the CEOs of Google, Facebook, and Twitter were asked to provide justification for their companies failure to curb the spread of such content, which led to incidents such as the Capitol riots. Topics discussed also included Facebook’s algorithms constantly promoting false information to keep up user engagement. Despite the “grilling” session, the Congress couldn’t come to any resolution for this problem. However, Facebook did try to “influence the decision” by recommending increased transparency reporting of such incidents. This is ironic as several critics have condemned Facebook’s transparency report as being biased and “not transparent enough”.
🏛️ Biden nominates tech critic Lina Khan to the FTC CNBC
Another Big Tech critic was recently nominated to the Biden administration, more hints about Biden’s mistrust in the Silicon Valley based tech companies.
CheckStep News
📣 CheckStep is now a Full Member of OSTIA, an industry body for UK-based organisations that develop solutions to improve online safety.
📣 Guillaume Bouchard, CheckStep’s CEO and Co-Founder was elected President and QCRI’s Preslav Nakov was elected as Secretary of the Conference for Truth and Trust Online.
📣 CheckStep is co-organising the Conference for Truth and Trust Online 2021. The call for papers has been issued inviting both technical papers and talk proposals from scholars, activists, developers, lawyers, ethics experts, fact-checkers, public servants, journalists, and researchers alike.
📣 CheckStep’s Deputy Head of Research Isabelle Augenstein and QCRI’s Preslav Nakov are amongst the guest editors working on a special issue of the Cambridge University Press Journal Natural Language Engineering, about hate speech detection on social media platforms.
📣 A new pre-print from the CheckStep research team is available now on arXiv: “A Neighbourhood Framework for Resource-Lean Content Flagging” presents an efficient, effective & inherently interpretable framework for content flagging.
📣 The CheckStep CheckPoint publication has a new article discussing the evolution of content moderation policies.
To know more about CheckStep, please click here.