Facebook isn't exactly enthusiastic about President Biden's claim that it and other social networks are "killing people" by allowing COVID-19 misinformation to spread. The social media firm posted a refutation of the allegations, using data to suggest that something other than Facebook was responsible for a slowdown in vaccination rates and a rise in cases.

The company noted that vaccine acceptance in user polling had risen from 70 percent in January 2021 to as high as 85 percent in July, and that cultural group disparities had declined "considerably" over the same period. This was ahead of Biden's goal of getting 70 percent of Americans vaccinated by July — to Facebook, this was a sign the company was "not the reason" the US fell short of that target.

Facebook added that Canada and the UK had higher vaccination percentages despite using the social network about as much as their American counterparts. There's "more than Facebook" to the US results, the company said. It also pointed to its efforts to both promote accurate claims and fight falsehoods, including the use of misinformation labels, reduced exposure and takedowns. 

The internet giant didn't attempt to find an alternate explanation for US troubles. Some observers have pointed to a possible link between political affiliation and vaccination rates, but Facebook didn't even hint at this in its refutal.

It's not a flawless argument. Facebook is trying to draw a link between its polling data and the entire US, which doesn't make for a neat and tidy comparison. The company also hasn't shared estimates of how much COVID-19 misinformation slips through the cracks. The social site has a strong incentive to downplay its possible contribution to the problem given past complaints that it hasn't done enough to stop misinformation campaigns. 

At the same time, the data shifts the attention back to the Biden administration — it may need to provide more substantial data if it's going to show that health misinformation on social networks like Facebook is a major threat, as the US Surgeon General recently said. If nothing else, it suggests the answer is a complicated one regardless of how much Facebook is responsible.

Joe Biden said that Facebook and other social media platforms are “killing people” by allowing misinformation about COVID-19 to spread on their platforms.

Biden’s comments came in response to a reporter who asked the president what his message to “platforms like Facebook” was regarding misinformation about COVID-19. “They’re killing people,” Biden said. “I mean they’re really — look, the only pandemic we have is among the unvaccinated. And they’re killing people.”

His remarks, one day after the Surgeon general issued an unusual health advisory on the dangers of vaccine misinformation, comes amid mounting pressure for Facebook and other platforms to do more to address misinformation about the coronavirus vaccines. But Facebook has come under particular scrutiny due to its size, and spotty history with countering vaccine falsehoods.

A widely cited reported from the Center for Countering Digital Hate found that much of the vaccine misinformation that spreads online can be linked to just 12 individuals — many of whom remain active on Facebook despite the company’s attempts to crack down on vaccine misinformation in recent months. Facebook didn’t immediately respond to a request for comment.

Last month, Instagram held its first-ever Creator Week, a virtual event the company described as “a life-changing three days with new feature news and celeb drop-ins.” One of those drop-ins was CEO Mark Zuckerberg, who made a brief appearance to share a message with creators.

“I think that any good vision of the future has to involve a lot more people being able to make a living by expressing their creativity and by doing things they want to do, rather than things they have to — and having the tools and the economy around them to support their work is critical,” he said. “Our goal is to be the best platform for creators like you to make a living.”

This week, Zuckerberg went even farther, announcing that Facebook plans to invest $1 billion in creators by the end of 2022. The investment will fund bonus programs, creator funds and other monetization programs to boost all stripes of creators on its platform.

That Facebook is funneling so much money and resources toward creators is indicative of not just the opportunity the company sees, but how much ground it has to make up.

For years, Facebook simply didn’t do much for creators. While Instagram has long had its own influencer community, the company has at times tried to limit their reach. Instagram’s founders were reportedly uncomfortable with the rise of influencers, and introduced an algorithmic feed to ensure users would see more posts from friends and family than brands and businesses.

While YouTube has offered monetization features for more than a decade, Instagram didn’t offer any kind of revenue sharing feature until last year. And many creators often felt at odds with Instagram. The company’s ever-changing algorithm fueled suspicions that it “shadowbans” or otherwise penalizes users who post too much or about the “wrong” topics.

“Facebook has been late to the game in terms of supporting the creative community in a meaningful way,” says Qianna Smith Bruneteau, founder of the American Influencer Council, a trade group representing the creator industry.

But Facebook is now trying to reverse those perceptions. For the past year, the company has been steadily churning out new tools for creators to make money. Since last May alone, the company has introduced a dizzying number of money-making features.

On Instagram, creators can now make money from commercials in IGTV or open their own shops. They can sell badges and products in live streams. On Facebook, they can host paid virtual events, promote fan subscriptions, or sell in-app gifts in live streams or audio rooms. Soon, they’ll be able to start paid newsletters, earn affiliate commission from products their followers buy and participate in a branded content marketplace. The company is also launching several new bonus programs that will pay creators for signing up for IGTV ads, creating Reels or meeting live-streaming milestones.

Facebook

Zuckerberg and other top executives now regularly speak about creators and the opportunity they represent. The company is so eager to win over the creator community it’s promised it won’t take a cut of their earnings until 2023.

Li Jin, founder of Atelier Ventures, a venture capital firm that invests in the creator economy, says surging interest in creators is because the industry has gotten so big it’s no longer something platforms can afford to ignore.

“I think for a long time there was no need to separately think of creators as a distinct segment that was in need of specialized features or funds,” Jin says. “I think what changed is the realization that … these creators’ content is driving a disproportionate amount of activity and engagement on the platforms.”

That Facebook is late to the creator economy also means the company is facing an incredible amount of competition. TikTok, which has a reputation for a creator-friendly algorithm, just passed 3 billion downloads, the first non-Facebook owned app to do so, according to analytics company Sensor Tower. Users of TikTok, and its Chinese counterpart Duoyin, together spent more than a half billion dollars in the app during the second quarter of 2021, alone. In the United States in 2020, TikTok was significantly ahead of Facebook and Instagram in user engagement, according to App Annie.

App Annie

Meanwhile Twitter, Snapchat, Pinterest and other platforms are also pouring money into new initiatives for creators. “There's a limited number of creators and everyone is in competition for them,” Jin says.

Facebook has offered various explanations for its sudden interest in creators. Zuckerberg has said he wants to help more people “make a living” off Facebook’s services. Instagram chief Adam Mosseri recently said the company was responding to “the shift in power from institutions to individuals across industries.”

It’s also a major opportunity to shift Facebook’s business away from ads. Though Facebook has promised it won’t take a cut of creators’ earnings for more than a year, that will eventually change (the company hasn’t said what its cut will be, only that it will be “less” than Apple’s 30-percent commission).

Creators could also provide a massive boost to the company’s push into shopping. Commerce has also been a major focus for the social network, which has already crammed shopping features into nearly every corner of Instagram, and Zuckerberg has said he intends to create “a full-featured commerce platform” across Facebook’s services.

What’s less clear is just how much creators will be willing to buy-in to Facebook’s vision. While a $1 billion investment will almost certainly fuel more interest in the platform, it’s not clear if it will prompt the kind of content Facebook might be hoping for. Instagram’s Reels, for example, was meant to be the company’s chief TikTok competitor. Yet the company has at times had to push creators to post original content there.

And concerns about Facebook’s algorithms remain, says Bruneteau. “The algorithm should be favorable to creators like it is on TikTok,” she says. “You have these instant influencers on TikTok, who have been able to grow million-plus followings in less than a year. However those same instant influencers who have those accounts have a tendency to have less followers on Instagram.”

There are signs that Facebook might be willing to address these concerns. Mosseri recently raised eyebrows when he said that Instagram is no longer a photo-sharing app, and that the company was working one ways to insert more recommended content in users’ feeds in order to compete with TikTok.

But even with a kinder algorithm, both Bruneteau and Jin caution that creators should be cautious in throwing too many resources into Facebook or any one platform.

“When creators are building their processes on top of these like centralized platforms, they're actually creating more value for the underlying platform than they're able to create for themselves,” Jin says. “At the end of the day you're strengthening Facebook's dominance because the more content you put there, the more it attracts consumer users and the more that translates into Facebook revenue and Facebook's network effects.”

Last April, Facebook’s AI research lab (FAIR) announced and released as open source its BlenderBot social chat app. While the neophyte AI immediately proved far less prone to racist outbursts than previous attempts, BlenderBot was not without its shortcomings. For one, the system had the recollection capacity of a goldfish — any subject or data point the AI wasn’t initially trained simply didn’t exist in its online reality, as evidenced by the OG BB’s continued insistence that Tom Brady still plays for the New England Patriots. For another, due to its limited knowledge of current events, the system had a strong tendency to hallucinate knowledge, like a digital Dunning-Kruger effect. But the advancements BlenderBot 2.0 displays, which FAIR debuted on Friday, should make the AI far more sociable, knowledgeable, and capable.

While BlenderBot 1.0 could only maintain its memory for a single discussion, its successor can remember topics of conversation over the course of multiple talks that can take days, weeks or even months to complete thanks to the implementation of a long-term memory module. What’s more, the AI can actively update its knowledge base by searching the internet for the latest news and details on any subject that the user wishes to speak about.

“BlenderBot 2 queries the Bing API for search results based on a generated search query, and conditions its response on the top few results,” Kurt Shuster, Research Engineer at Facebook AI, told Engadget. “We rely on Bing to provide high quality search results.” As such, BlenderBot 2.0 is now capable of speaking coherently about breaking news and new media, not just the data it was trained upon.

FAIR

“BlenderBot 2 is limited only by what a powerful search engine can provide,” Jason Weston, Research Scientist at Facebook AI, added. So for example, if you are more interested in learning about Tom Yewcic (the Patriot’s combo QB/Punter from the 1962 season) than you are about Tom Brady, BB 2.0 has you covered. It’s the same with more scholarly subjects, like photosynthesis or redox reactions, Weston continued. So long as the information is available on the web, “there is no reason BlenderBot 2 cannot discuss this.”

By actively searching the internet for information, BlenderBot 2.0 can also reduce the instances in which it hallucinates knowledge. “Providing the system with more commonsense reasoning will allow BlenderBot to make sure it does not confuse subtle concepts,” Weston explained, “such as a movie director versus a producer or a pitching coach versus a hitting coach.”

FAIR

The only wrinkle really occurs when discussing non-english based media, such as Demon Slayer: Kimetsu no Yaiba. “It is reasonable to conclude Bing will surface information about it and BlenderBot 2 can use that information accordingly,” he said. “We currently focus on english-based search results, so non-english references may not be fully covered.” The system will, however, recognize that Demon Slayer is of interest to you and will be more likely to bring up manga-centric subjects in future discussions.

FAIR has taken multiple steps to ensure that BlenderBot does not become the next Tay. “BlenderBot 2 does not learn directly from user input, as Tay did,” Shuster said. “We have taken extensive safety steps to ensure that BlenderBot 2 can handle adversarial users. Specifically, we employ both baked-in and two-stage techniques. BlenderBot 2 can detect itself if the incoming context will result in an offensive response, and additional safety layers where a safety classifier can detect if either the user input or the bot's output is offensive. Each handles the response appropriately.”

And while the system is currently focused on chewing its way through the English language corpus, FAIR does see BlenderBot does eventually extend to other languages as well. “While not in our immediate plans, the goal of our team is to build a superhuman conversationalist,” Shuster said. “This kind of agent requires multilingual understanding.”

Recent internal benchmarking processes found that BlenderBot 2.0 outperformed its predecessor by 17 percent in its engagingness score and 55 percent in its use of previous conversation sessions according to human evaluators, per a Friday blog from FAIR. What’s more, BlenderBot's rate of knowledge hallucinations dropped from 9 percent (!) in BB 1.0 down to just 3 percent in the current iteration.

Looking ahead, “humans interacting with AI systems via discourse is the future of AI,” Weston asserted, “and ensuring that humans have an engaging, informative experience is critical to that future. BlenderBot 2 combines the engagingness of BlenderBot 1.0 with the knowledge capabilities of a system with access to the entire internet, so ostensibly we are on the right track.”

Twitter rolled out voice tweets over a year ago now and has taken a lot of heat for the lack of accessibility features. Now, it's finally rolling out automatically generated captions that appear when you click on the "CC" button. The new feature is only available on iOS, as voice tweets have yet to arrive on Android. 

Captions are available in English, Japanese, Portuguese, Turkish, Arabic, Hindi, French, Indonesian, korean and Italian. They will only appear on new voice tweets, as they need to be generated when the tweet is created, Twitter told The Verge. 

When voice tweets were first being tested last June, critics immediately pointed out that they should have had captions from the start as required in the US by Federal Law. Twitter then admitted that it didn't have a dedicated accessibility team and relied on employees to donate additional time for those features. Since then, however, the company has launched teams dedicated to accessibility. The company originally promised to add automated captions by early 2021, but that date obviously slipped a bit. 

Twitter promised to improve and expand the service across its products. “Though it’s still early and we know it won’t be perfect at first, it’s one of many steps we’re taking to expand and strengthen accessibility across our service, and we look forward to continuing our journey to create a truly inclusive service," said Twitter's global accessibility head Gurpreet Kaur. 

Facebook is testing a change that will let users know when their post was removed as a result of automation. The new experiment comes in response to the Oversight Board, which has said the social network should be more transparent with users about how their posts are removed.

The company revealed the new test in a new report that provides updates on how Facebook is handling the Oversight Board’s policy recommendations. The test comes in response to one of the first cases the Oversight Board took up, which dealt with an Instagram post meant to raise awareness of breast cancer that the company removed under its nudity rules.

Facebook restored the post, saying its automated systems had made a mistake, and updated Instagram’s rules to allow for “health-related nudity.” But the Oversight Board had also recommended that Facebook alert users in cases when a post was removed with automation rather than as a result of a human content reviewer. Facebook previously said it would test this change, which is now in effect.

“We’ve launched a test on Facebook to assess the impact of telling people more about whether automation was involved in enforcement,” facebook writes in its report. “People in the test now see whether technology or a Facebook content reviewer made the enforcement decision about their content. We will analyze the results to see if people had a clearer understanding of who removed their content, while also watching for a potential rise in recidivism and appeals rates.” The company added that it will provide an update on the test later this year.

The report also shed some additional insight into how the company is working with the Oversight Board. The report notes that between November 2020 and March 2021 it referred 26 cases to the board, though it’s only chosen to take up three — one of which was in response to its suspension of Donald Trump. (Notably, the latest report only covers the first quarter of 2021, so it doesn’t address the board’s recommendations in response to Trump’s suspension.)

Though the Oversight Board has only weighed in on a handful of cases, its decisions have resulted in a few policychanges by Facebook that could have a much broader effect. However, in some areas, the company has declined to follow up on its policy suggestions, such as one that Facebook study its own role in enabling the events of January 6th. In a blog post, the company noted that “the size and scope of the board’s recommendations go beyond the policy guidance that we first anticipated when we set up the board, and several require multi-month or multi-year investments.”

Instagram is running a new test to tell users about another app they might want to check out: Facebook. The photo sharing app is experimenting with a notice at the top of users’ feeds that encourages them to check out features that are “only available” on Facebook.

“We’re testing a way to let people who have connected their Instagram accounts to Facebook know about features only available there, such as how to find a job, date online, buy and sell goods, or catch up on the latest news,” a spokesperson said in a statement.

That the company is using one of its billion-user apps to promote another billion-user app might not seem to make a lot of sense, but it’s only the latest (and perhaps most aggressive) way the social network has used Instagram to drive people back to its main app. The company has been steadily bringing the two apps closer together and has been encouraging users to link their accounts. (A book published last year reported that Mark Zuckerberg was “jealous” of Instagram’s success and worried the app could eventually “cannibalize” Facebook. Tensions between him and the app’s founders ultimately led to their departure in 2018.)

Facebook points out that only “a very small group” of Instagram users who have previously opted to link their accounts will see the messages, which can be dismissed. But even if it never expands, it would suggest that the company is far from done with its attempts to get Instagram users to spend more time on Facebook.

On Thursday, Facebook disclosed that a network of hackers with ties to Iran tried to use its platform to target US military personnel. At the center of the campaign was a group known as Tortiseshell. Facebook says the collective went after individuals and companies in the defense and aerospace industries. Its primary targets were in the US, but they also sought out people in the UK and parts of Europe.

“This activity had the hallmarks of a well-resourced and persistent operation, while relying on relatively strong operational security measures to hide who’s behind it,” Facebook said. "Our platform was one of the elements of the much broader cross-platform cyber-espionage operation, and its activity on Facebook manifested primarily in social engineering and driving people off-platform (e.g. email, messaging and collaboration services and websites), rather than directly sharing of the malware itself."

What went down appears to be unprecedented for Tortoiseshell. In the past, the group has primarily targeted IT companies throughout the Middle East. The methods it employed were similar to those that China’s Evil Eye used to target the Uyghur community earlier in the year.

Facebook says the group created “sophisticated online personas” to contact its targets and build trust with them before trying to convince them to click on malicious links. They had accounts across multiple social media platforms to make their ruse appear more credible. The group built fake recruiting websites and even went so far as to spoof a legitimate US Department of Labor job search tool. Facebook believes at least some of the malware the group deployed was developed by Mahak Rayan Afraz, a company with ties to the Islamic Revolutionary Guard Corps.

Iran has been accused of a variety of malicious online activities over the past year. Most notably, Microsoft said last September it was one of the countries that tried to meddle in the 2020 US presidential election.

Almost two years down the line, Facebook is taking a big step to expand its payments platform. The company has announced plans to break out Facebook Pay from its own ecosystem to facilitate transactions across the web. Starting in August, US-based customers will be able to make purchases from Shopify-powered businesses using the payment method. When you visit a supported seller, you'll see a Facebook Pay button allowing you to complete a purchase with your saved card or PayPal details. 

In that sense, Facebook will be jostling for attention at checkout with rival mobile and digital wallets including Apple Pay and Google Pay. Of course, each has a built-in audience based on preferred hardware, operating system or, in Facebook's case, social network. 

Until now, Facebook Pay has been limited to the company's own platforms including Messenger, Instagram and WhatsApp. Facebook has pitched it as an expedient way to shop on those apps, split the bill with friends, send money abroad and donate to charities. Expanding Facebook Pay's reach across the web not only boosts its convenience, but also fits into the company's wider ecommerce strategy. In June, Facebook added support for QR codes to the payments platform, enabling users to send money to people outside of their friends group. 

At the same time, it has integrated shopping features into its wider ecosystem of social apps to capitalize on the shift to ecommerce. In May, Facebook introduced Shops that turn business pages on its main platform and Instagram into online storefronts. Like Google and Snapchat before it, the company is also turning to visual search to help people discover more shoppable products on Instagram. Not to be left out, WhatsApp also received a shopping button that let users browse a retailer's product catalog. 

Twitter is killing Fleets less than a year after launching the Stories-like feature to all its users. All Fleets will disappear for the final time on August 3rd.

The short lived feature was at times controversial. Soon after it rolled out to all Twitter users last November, many raised questions about how the feature could be used to target others for harassment. Others questioned whether Twitter really needed a “Stories” feature of its own.

In a blog post, Twitter VP of Product Ilya Brown said the company hadn’t “seen an increase in the number of new people joining the conversation with Fleets like we hoped.” Brown added that Spaces will continue to get placement at the top of users’ timelines and that the company is still analyzing the full-screen ads it started testing in Fleets last month (Twitter hinted at the time that the new ad format could eventually make its way to other places in the service, too).

Both Brown and Twitter Product Lead Kayvon Beykpour pointed out that “winding down features every once in awhile” is something the company fully anticipates as it tries to reinvent itself. Twitter has been working on a number of new features that could dramatically change its service – including subscriptions and paid features for creators — and has publicly previewed several newideas it’s considering in recent months.