Facebook’s “state-controlled media” labels appear to reduce engagement with content from authoritarian nations. A new study reveals that, with the added tags, users’ engagement decreased when they noticed content labeled as originating from Chinese and Russian government-run media. However, the labels also appeared to boost user favorability of posts from Canadian state media, suggesting broader perceptions of the country play into the tags’ effectiveness.

Researchers with Carnegie Mellon University, Indiana University and the University of Texas at Austin conducted the set of studies which “explored the causal impact of these labels on users’ intentions to engage with Facebook content.” When users noticed the label, they tended to reduce their engagement with it when it was a country they perceived negatively.

The first experiment studied 1,200 people with US-based Facebook accounts — with and without state-controlled media labels. Although their engagement with posts originating from Russia and China went down, it only had that effect if they “actively noticed the label.” A second test in the series observed 2,000 US Facebook users to determine that their behavior was “tied to public sentiment toward the country listed on the label.” In other words, they responded positively to media labeled as Canadian state-controlled and negatively toward Chinese and Russian government-run content.

Meta

Finally, a third experiment examined how broadly Facebook users interacted with state-controlled media before and after the platform added the labels. They concluded the change had a “significant effect” as the sharing of labeled posts dropped by 34 percent after the shift, and user likes of tagged posts fell by 46 percent. The paper’s authors also noted that training users on the labels (“notifying them of their presence and testing them on their meaning”) significantly boosted their odds of noticing them.

“Our three studies suggest that state-controlled media labels reduced the spread of misinformation and propaganda on Facebook, depending on which countries were labelled,” Patricia L. Moravec, the study’s lead, wrote in the paper’s summary.

However, the studies ran into some limitations in determining correlation vs. causation. The authors say they couldn’t fully verify whether their results were caused by the labels or Facebook’s nontransparent newsfeed algorithms, which downlink labeled posts (and make related third-party research exceedingly difficult in broader terms). The paper’s authors also note that the experiments measured online users’ “beliefs, intentions to share, and intentions to like pages” but not their actual behavior.

The researchers (unsurprisingly, given the results) recommend social companies “clearly alert and inform users of labeling policy changes, explain what they mean, and display the labels in ways that users notice.”

As the world grapples with online misinformation and propaganda, the study’s leads urge Facebook and other social platforms to do more. “Although efforts are being made to reduce the spread of misinformation on social media platforms, efforts to reduce the influence of propaganda may be less successful,” suggests co-author Nicholas Wolczynski. “Given that Facebook debuted the new labels quietly without informing users, many likely did not notice the labels, reducing their efficacy dramatically.”

Meta will soon offer Stories and Reels in chronological order, among other changes, to comply with the European's Digital Services Act (DSA), the company announced. The changes were expected after the European Commission announced that it had reached an agreement in April to create new rules that would require platforms like Facebook to offer alternative systems "not based on profiling" as a key requirement. 

Meta said it has mobilized over 1,000 people to "develop solutions to the DSA's requirements." Some of the changes will increase transparency about how its systems work and provide users more options to tailor their experiences on Facebook and Instagram. At the same time, it's establishing an "independent compliance function" to ensure it meets ongoing regulatory obligations. 

Starting later this month, Meta will offer Reels, Stories, Search and other parts of Facebook and Instagram that are unranked by Meta using its AI recommendation process. "For example, on Facebook and Instagram, users will have the option to view Stories and Reels only from people they follow, ranked in chronological order, newest to oldest," wrote Meta President of Global Affairs Nick Clegg. 

It's not clear how Meta will implement the change. The main Feed on Instagram already allows users to sort by Following instead of using the algorithm-based approach. However, the "Following" feature is effectively a secondary page on Instagram, and the app always defaults to the algorithmic "For You" option when first opened. Facebook is even more of a hassle, forcing you to select a menu, go into Feeds and tap "Friends" rather than "All." 

Users will also be able to view Search results based only on the words they enter, rather than results personalized specifically to them based on their previous activity and personal interests. The company is also providing more information about how its AI systems rank content via 22 system cards for Facebook and Instagram, adding to its "Why Am I Seeing This" feature. 

"These cards provide information about how our AI systems rank content for Feed, Reels, Stories, and other surfaces; some of the predictions each system makes to determine what content might be most relevant to people; and the options available to help customize an experience on Facebook and Instagram," Meta said.

Meta is also expanding its Ad library to display and archive all ads (for one year) that target EU users, including date run, parameters used for targeting (age, gender, location), who received the ad and more. It's also rolling out two new tools for researchers that include publicly available content from Pages, Posts, Groups and Events. 

The company said that it "welcomes the principles of transparency, accountability and user empowerment at the heart of the DSA," adding it has "long advocated for a harmonized regulatory regime." However, Meta previously expressed extreme displeasure when Apple introduced changes that allowed users to easily opt out of targeted advertising starting with iOS 14. To that end, observers will no doubt be keenly interested in how the changes are implemented and whether they follow the letter, if not the spirit, of the new law.

YouTube is quickly becoming a home for AI-generated music, and the service is trying to strike a balance between the technology's fans and the labels eager to protect their copyrights. The company and partners like Universal Music Group (UMG) have unveiled a set of principles for AI music. In theory, the approach encourages adoption while keeping artists paid.

To start, YouTube maintains that "AI is here" and that it must have a "responsible" strategy. Accordingly, it's forming a Music AI Incubator that will influence the company's strategy. UMG and artists it represents (including Rosanne Cash, Yo Gotti and Frank Sinatra's estate) will help gather insights from YouTube's AI experiments.

YouTube also says AI music must include "appropriate protections" against copyright violations, and must also provide "opportunities" for partners who want to get involved. While the video giant hasn't detailed what this will entail, it suggests it will build on the Content ID system that helps rights holders flag their material. On top of this, YouTube claims it will scale its content policies and safety structure to adapt to AI. The firm already has systems in place to catch copyright abuse, misinformation and other violations, but intends to pour more resources into those methods.

The principles are currently vague and don't do much to change YouTube's stance. More details are due in the months ahead, however, including policies, particular technologies and monetization for creators.

Generative AI is increasingly popular for unauthorized collaborations and mashups (including for UMG artists like Drake and Frank Sinatra), but it's also finding legitimate uses. The surviving members of The Beatles are using AI to create a 'final' song from a John Lennon recording, while electronic artist Holly Herndon covered Dolly Parton using an AI voice. UMG itself is exploring AI-made soundscapes. YouTube's principles could help it profit from legal productions while dodging lawsuits from artists and labels worried about ripoffs.

Earlier this year, Pew Research reported that a majority of US Twitter users reported spending less time on the platform since Elon Musk’s takeover last year. Now, new data suggests another important group of users are also pulling back from the service now called X.

More than half of scientific researchers who use Twitter report they’ve reduced the amount of time they spend there or have left altogether, according to a survey of thousands of scientists conducted by Nature. And nearly half of those polled said they’ve turned to alternative social networks like Mastodon

Of the 9,200 researchers polled, more than 47 percent said they had decreased their usage of the site, while nearly 7 percent reported quitting the site altogether. Notably, almost the same number said they had started an account on at least one new platform over the last year.

Of these, Mastodon, which has seen significant growth since Musk’s takeover of Twitter was announced, was the most widely used. About 47 percent of researchers said they had started using the open-source platform in the past year. LinkedIn and Instagram were the next most popular, drawing 35 and 27 percent of researchers, respectively. Interestingly, Meta’s Twitter competitor, Threads, took the number four spot even though the app launched only days before Nature conducted the poll.

As with the earlier data from Pew, Nature’s findings suggest that Twitter usage is down among those who were once active on the platform. It also highlights how much the dynamics of Twitter have changed over the last year.

Twitter, as Nature points out, has historically been an important platform for researchers and scientists. It’s been used to publicize research and promote scientific debate. And Twitter’s researchers have served as an important source of authoritative information on a platform that's long struggled to combat misinformation. Twitter has also been a valuable source of data for countless researchers studying everything from public health to linguistics.

But much of that has now changed. Many users now feel that their voices are drowned out on a platform that prioritizes content from those with paid verification. And the company has made its API for researchers so expensive that most can no longer access it. So while not all the researchers that spoke to Nature were ready to give up on Twitter entirely, it does seem the company’s tactics have alienated large swaths of the scientific community.

X didn't respond to a request for comment. 

Despite having trouble obtaining former President Donald Trump's records from Twitter (now X), the special counsel investigating the events of January 6th, 2020 was ultimately able to get his hands on vast amounts of information from the website. Based on the newly unsealed court filings (PDF) shared by Politico, though, it took a a lot of back and forth and a massive amount of effort before the counsel was able to secure the data requested from Twitter. 

Apparently the search warrant required Twitter to produce all the tweets the @realdonaldtrump created, drafted, liked or retweeted, including any post that he eventually deleted. The website also had to turn over DMs "sent from, received by, stored in draft form in, or otherwise associated with" @realdonaldtrump, as well as all other accounts related to it that the former president might have used on the same device. All devices used to log in to @realdonaldtrump had to be noted and given to the counsel, as well, along with the account's privacy settings and IP address history from October 2020 to January 2021. In addition, Twitter was required to hand over all records of searches done by @realdonaldtrump and the account's location information from the same time period.

Throughout the transcript of the proceedings, you'll see how the prosecutors insisted that Twitter had to adhere to a nondisclosure order that prohibits the company from telling Trump about the warrant. They explained that they had evidence that disclosing the warrant could jeopardize the probe, warning the court that "there would be actual harm and concern for the investigation, for the witnesses going forward." Meanwhile, Twitter's side argued that some of the information the investigators were requesting could be covered by executive privilege, which the prosecutors and judge questioned, since they didn't deem it likely that Trump discussed official government matter via DMs. 

As Politico reports, US District Judge Beryl Howell notably called attention to Twitter's efforts to give Trump advance notice about the search warrant. She had called the action "extraordinary" and something the company had never done for other users. The judge wondered whether Elon Musk was trying to "cozy up" to the former president and make him feel "particularly welcomed" on the platform. Trump's Twitter account had already been reinstated after Musk took over the website, but he has yet to start posting again. 

Telegram launched its Stories feature to everyone today, following its availability to Premium users starting last month. Like Facebook Messenger’s Stories, they appear as expandable bubbles above your conversation. However, Telegram’s take is more customizable, providing granular control of who sees uploaded posts and for how long. “Now when you meet people on Telegram, you’ll see exciting snapshots of their life — not just a few profile photos,” the company wrote in a blog post today.

Launching alongside the messaging service’s 10th birthday, Telegram describes Stories as “by far the most-requested feature” in the company’s decade-long history. Its privacy controls include visibility options for everyone, all contacts, selected contacts or close friends.

Telegram

Telegram’s Stories also let you hide your posts from contacts you don’t want to see, and Premium users can choose between six, 12, 24 and 48 hours of visibility for new posts. In addition, post creators can see a list of the Telegram users who viewed their content. It also supports a BeReal-like dual-camera mode, letting you simultaneously share photos or videos captured by your phone’s front and rear sensors. The feature also includes reactions, so viewers can add a heart or choose from “hundreds” of other responses to posts.

Some of Stories’ more advanced controls are reserved for Premium subscribers ($5 per month). Perhaps most significant, paying users’ posts display first, giving them more exposure. Subscribers can also view others’ stories in stealth mode, hiding all traces of their visit from the author. Additionally, subscribers get the previously mentioned custom expiration options, a permanent view history (see who viewed your posts even after they expire), the ability to save Stories to the gallery, “10 times longer” captions and a higher allotment of daily Stories (up to 100).

Telegram Stories is scheduled to roll out to the service’s iOS and Android apps today.

Instagram now lets you add music to photo carousels. Unveiled in partnership with pop star Olivia Rodrigo to promote her single “bad idea right?”, the feature allows you to pick licensed music to soundtrack your slideshows. In addition, the company announced that you can create Collabs with up to three co-authors and post audience-response prompts to Reels.

The carousel soundtracking feature adds a missing piece already found in TikTok’s Photo Mode, launched last year. “Whether you’re sharing a collection of summer memories with friends or moments from your camera roll, you can now add music to your photo carousels,” Instagram wrote in a blog post today. “Building off our launch of music for feed photos, anyone can add a song to capture the mood and bring their carousel to life.”

Also announced today, Instagram Collabs adds the ability to invite up to three friends (up from one) to help co-author feed posts, carousels or reels. The platform says each contributor’s audience will see the content (perhaps hinting that it could be a handy way for influencers to benefit from each other’s followings) and will feature on each account’s profile grid. In addition, the company says private profiles can still start posts / reels and invite collaborators as long as they follow the private account.

Instagram

Instagram also updated how the Add Yours sticker works. When a creator adds the new Add Yours prompt to a Reel and followers contribute content as a response, the creator can now highlight their favorite posted replies for all their followers to see. “With the Add Yours sticker, a creator or artist can invite their followers to join in on a fun prompt or challenge they create on Reels, and then hand-pick their favorite submissions to celebrate their fans’ creativity.” It essentially sounds like a way to use the human social desire to connect with high-status figures (especially celebrities like Rodrigo) to build engagement for creators and the platform as a whole.

Finally, Instagram noted that it’s bringing its music library “to more countries over the coming weeks,” although it hasn’t yet announced specific nations or dates. However, it did mention that Instagram is partnering with Spotify in Mexico and Brazil to showcase 50 of the most popular songs on Instagram Reels on the music platform’s Reels Music Chart.

X, the social network formerly known as Twitter, is now letting users download videos from its website, as long as a couple of criteria are met. First, the creator must explicitly allow downloads for the content they post. Also, the ability to save videos from the website is exclusively available to verified (aka paying) users. Elon Musk has announced the new feature on his account and explained that a Blue subscriber can download content by tapping the ellipsis (...) menu on the upper right corner of the screen when a video is in full screen mode. Musk also said that X will soon make it easy to download videos on mobile simply by tapping and holding the content on the screen. 

Previously, people have had to use to third-party websites and apps to be able to save videos. Of course, non-paying users still have to use those tools if they want to download content from X. That doesn't come as a surprise, seeing as the company has been releasing new features specifically for paying members while making the website less usable for the rest of its userbase. It announced in July that it will limit the number of DMs non-paying users can send to address its spam issues, and it also previously limited the number of posts non-verified users can see to 600 a day.

Meanwhile, X gave verified users the ability to publish posts with up to 25,000 characters and allowed them to post videos up to three hours in length. More recently, it also gave Blue subscribers the option to hide their blue checkmark in case they don't want to make it obvious that they're paying for the service. For those interested to see how X has evolved since Elon Musk took over Twitter, company CEO Linda Yaccarino has just published the list of changes it has implemented right here.

One of the main selling points for Twitter Blue — now "X Blue" — when the service was first launched was that anybody on the platform willing to pay for it can get the once-coveted blue checkmark. Over the past months, though, subscribers have been getting shamed for paying $8 a month or $84 a year for the service. Now as TechCrunch has noticed, the company has updated its support page for X Blue with a new feature for members: The ability to hide the verified checkmark on their account. 

Under the "Profile customization" section in account settings, subscribers will now find a new "Hide your blue checkmark" option that they can tick. By activating the feature, the badge will no longer show up on their profiles and next to their usernames on posts. However, the company warned that it could still show up in some places and that some features may not be available to them while their checkmark is hidden. It didn't say which features will become inaccessible, but Twitter has rolled out a number of changes made specifically for paying users since Elon Musk took over. 

It has increased paid users' post limit to 25,000 characters and is even working on new tools to publish long-form content. The website has also expanded their video limit to 3 hours. Meanwhile, the social network has become less and less attractive for free users. It announced in July that it will limit the number of DMs non-paying users can send in an effort to limit spam, and it previously put a strict cap on how many tweets a day a user can see due to "extreme levels of data scraping." While the restriction was temporary, unverified accounts were initially limited to 600 posts daily. 

App developer Alessandro Paluzzi first spotted the capability to hide checkmarks in March. Based on the screenshots Paluzzi shared, the verification process will remain the same, with users being required to submit a government ID to authenticate their identities. It now simply won't be obvious at first glance that someone's paying for X Blue.

TweetDeck has started showing signs that it will not escape Twitter's massive rebranding unscathed. If you take a look at the tool's landing page while logged out, you'll see "XPro" on the upper left corner of the website, indicating its new name. That's pretty much it at the moment, though — the page still shows the iconic Twitter bird logo, and it still calls TweetDeck a "powerful, real-time tool for people who live on Twitter." And, yes, its URL is still https://tweetdeck.twitter.com/ in the same way Twitter's (sorry, we mean X's) URL is https://twitter.com/. 

Twitter's rebranding to "X" hit us overnight quite literally in July, with Elon Musk announcing the change in a tweet and then, reportedly, emailing his employees about it afterward. The social network started implementing changes after that by swapping out the Twitter bird with a generic-looking "X" logo, which will apparently "evolve over time," as well as renaming tweets to "posts" and retweets to "reposts." Google and Apple both relisted the social network's app as "X" on the Play Store and the App Store, respectively, by the end of July. 

The move, while sudden, perhaps didn't come as a surprise for those who've followed Musk's career over the years. Musk is known for favoring the "X" branding and had wanted to rename PayPal to X.com before he was ousted as CEO. The executive also renamed Twitter's parent company to X Corp. He said the social network's new name will help it branch out and become an "everything app" with payments and banking features. But critics are doubting the rebranding's success, with some calling it "the dumbest thing [Musk has] done since taking over" and something that now makes the company less identifiable and less valuable. 

Despite the criticisms, the erasure of the Twitter brand is now well underway. TweetDeck will not remain TweetDeck for very long, and it also won't be free to use in the near future. The company has previously noted that it will be locked behind a paywall and will be exclusively available to Blue subscribers.