Are Americans dying from Misinformation? š§
The Checkstep Round-Up is a monthly newsletter that gives you fresh insights for content moderation, combating disinformation, fact-checking, and promoting free expression online. The editors of the newsletter are Kyle Dent and Vibha Nayak. Feel free to reach out!
Itās been hard to keep up with Twitterās back-and-forths this month. First NPR and other publicly funded media outlets were labeled state-media, then they werenāt, and now (we think) there are no labels, but weāre not sure at this point. Then there is the matter of the blue check marksā¦ sorry, canāt help you; we donāt know whatās going on. Apart from Twitter, weāve got Podcasts! Depending on who you talk to, theyāre either a great source of news or the last bastion of unchecked disinformation or most likely both, so, choose carefully. Oh, and by the way, weāre also dying earlier, at least in part because of misinformation, so really choose carefully. Have fun out there!
Checkstep News
š£ Our CEO, Guillaume Bouchard, was featured on the podcast - AI Unleashed where he talks about everything from his story behind setting up Bloomsbury AI and Checkstep to the early signs of generative models in academia and the challenges that came with scaling them. Tune in to hear more!
š£ Guillaume was also interviewed by the Center for Data Innovation where he spoke about the use of AI in content moderation, the importance of explainability, and emergent risks.
š£ Part of the Checkstep team also attended the Trust & Safety Forum in Lille, where Kieron Maddison, our Product Manager, spoke about the operational consequences of the Digital Services Act.
Moderating the Marketplace of Ideas
š·ļø Twitter drops 'government-funded' label on media accounts, including in China (Reuters)
After weeks of a strange about-face, Twitter is no longer labeling certain media companies as state-media or publicly funded. Twitter had started labeling NPR a state-affiliated media organization putting it in the same category as Russiaās RT and Chinaās Xinhau. The designation was normally used for outlets where the state exercises control over their content. The label serves as a warning to people about the trustworthiness of the accountās content, and according to Twitter, content with the designation is not eligible for recommendation or amplification of any kind. For its part, NPR said that they will not tweet from the @NPR account until Twitter removes the false label, and have since declared that they have left Twitter entirely.
š Twitter is facing an exodus from its most devoted followers: Journalists (Press Gazette)
While Twitter lags behind most other social media platforms in general popularity, among journalists, itās considered the place to be. However, since Elon Muskās takeover of the platform about half have said they have already or are considering leaving.Ā
šŖļø Twitterās new limits may block your next tornado warning (The Washington Post)
Twitter is unlikely to remain an essential source for news for other reasons as well. The National Weather Service and broadcast meteorologists say new policies at Twitter will limit their ability to quickly alert the public to extreme weather changes. They recommend having multiple ways to receive weather information and alerts.
šµ Research: How People Feel About Paying for Social Media (Harvard Business Review)
Itās been a long free ride, but it might be coming to an end. Several social media platforms have started pay-for-play subscriptions. In a recent study, Harvard researchers asked more than 1,000 social media users how they feel about the new approach and how they plan to engage. Respondents said they were most likely to pay for Snapchat, then Instagram, Facebook and Twitter. Both college-educated users and politically conservative users were more open to a subscription approach but expected the features to be better.Ā
š§ Podcasts as a Source of News and Information (Pew Research Center)
Podcasts have become very popular among American audiences, especially younger ones. About half of Americans have listened to one in the past year and about 20% listen to at least a few every week. Itās mostly for entertainment and learning but most podcast listeners tune in for news as well. Butā¦
š¤Ø How podcasts have become misinformation machines ā and what can be done about it - National (Global News)
As podcasts are becoming more popular, they are also becoming a more frequent source of misinformation. Podcasts also pose difficulties for fact-checkers and content moderators, so theyāre becoming a favorite vector for those profiting from and pushing ideologies based on misinformation.
š± Post, a publisher-focused Twitter alternative, launches to public (TechCrunch)
A new alternative to Twitter with a plan to monetize launched in beta this month. Publishers can use the platform to charge users micropayments to read individual news items. Content might include articles from traditional media outlets and other types of media, like subscription newsletters, or even free and ad-supported content thatās offered elsewhere. The CEO of the new platform, Noam Bardin, was previously CEO of Waze at Google.
š©ŗ Health misinformation is lowering U.S. life expectancy, FDA Commissioner Robert Califf says (CNBC)
The U.S. has lower life expectancy than the average of high-income countries and misinformation is at least part of the reason according to the FDA Commissioner. Itās not just COVID-19 related misinformation either, and itās getting worse not better. The decline is attributed to choices people make based on things they read that have influenced their thinking.
š Misinformation Defense Worked in 2020, Up to a Point, Study Finds (The New York Times)
Efforts to limit the effects of misinformation following the 2016 election seemed to have worked to some extent. By the 2020 election many fewer voters had visited websites with false or misleading information according to a study from Stanfordās Social Media Lab. Researchers report that content labels and media literacy training were contributors to the decline. Despite this success, nearly 68 million visited sites that were not credible leading into the 2020 electionā¦
š·šŗ š½ Russians boasted that just 1% of fake social profiles are caught, leak shows (The Washington Post)
On the other hand, among the items discovered in the widely publicized intelligence leak on DiscordĀ is the fact that Russia is better at manipulating social media and search rankings than we might have suspected. Russian operators of fake accounts, spreading lies about Ukraineās military and side effects of vaccines, claim only 1% of their bogus accounts are detected.
š®š© This citizen-run organization is teaching thousands of Indonesians to fact-check (Nieman Lab)
Mafindo (an acronym for the Indonesian Anti-Slander Society) is combating misinformation and promoting media literacy in Indonesia through a team of nine people and thousands of volunteers across Indonesia who are conducting trainings in fact-checking and trying to get more people connected with their work.
šļø Supreme Court takes up disputes involving public officials who block critics on social media (CBS News)
This First Amendment issue initially came up because Trump was prone to blocking his critics on Twitter, but it never got settled since he both left office and was booted off Twitter. However, following his lead other public officials have been blocking their detractors on social media bringing the issue to the Supreme Court next term after all. The decision should come out sometime in the summer of 2024.
š¤ Will AI-generated images create a new crisis for fact-checkers? Experts are not so sure (The Reuters Institute for the Study of Journalism)
Gretel Kahn, from the Reuters Institute for the Study of Journalism, discusses with several journalists, experts, and fact-checkers the fast-moving situation with AI-generated fakes to determine the dangers and what we can expect as well as how the rapid production of fake imagery might affect fact-checking and content moderation.
*Donāt trust the music you listen to either.
Regulatory News and Update
š®š³ India to require Facebook and Twitter rely on govāt fact checking (TechCrunch)
India has modified its laws to prohibit social media firms from hosting any misleading content about essentially anything related to the government. The platforms are also now required to use the governmentās own fact-checking unit to determine if any claim is true or false. Big Tech, not surprisingly, is not a fan of the new laws.
šŗšø Montana on cusp of becoming first state to block TikTok downloads (NBC News)
The governor of Montana is expected to sign a bill that would block new downloads of TikTok making it illegal for app stores to make it available and also illegal for the company to operate within the state. People who already have TikTok installed are not prevented from using the app. We can expect this one to go to the courts soon.
Reading Corner
š” Hereās an article by Kieron, our Product Manager, on āWhen is it Appropriate to Read Private Messages on Educational Platforms?ā