This is default featured slide 1 title

Go to Blogger edit html and find these sentences.Now replace these sentences with your own descriptions.This theme is Bloggerized by Lasantha Bandara - Premiumbloggertemplates.com.

This is default featured slide 2 title

Go to Blogger edit html and find these sentences.Now replace these sentences with your own descriptions.This theme is Bloggerized by Lasantha Bandara - Premiumbloggertemplates.com.

This is default featured slide 3 title

Go to Blogger edit html and find these sentences.Now replace these sentences with your own descriptions.This theme is Bloggerized by Lasantha Bandara - Premiumbloggertemplates.com.

This is default featured slide 4 title

Go to Blogger edit html and find these sentences.Now replace these sentences with your own descriptions.This theme is Bloggerized by Lasantha Bandara - Premiumbloggertemplates.com.

This is default featured slide 5 title

Go to Blogger edit html and find these sentences.Now replace these sentences with your own descriptions.This theme is Bloggerized by Lasantha Bandara - Premiumbloggertemplates.com.

ByteDance CFO assumes role as new TikTok CEO

Eight months after former TikTok CEO Kevin Mayer quit in the midst of a full-court press from the Trump administration against the Chinese-owned social media giant, TikTok finally has a new permanent leader.

ByteDance’s recently hired CFO Shouzi Chew will be assuming the role as TikTok CEO while still holding the CFO role at its parent organization, the company announced Friday morning. It’s a bold move, likely signaling that the company believes that the worst of its tussles with the U.S. executive branch are over as President Biden has seemed uninterested in picking up former President Trump’s pet project.

Vanessa Pappas, who was serving as interim CEO, will take the role of COO going forward.

“The leadership team of Shou and Vanessa sets the stage for sustained growth,” ByteDance CEO Yiming Zhang said in a press release. “Shou brings deep knowledge of the company and industry, having led a team that was among our earliest investors, and having worked in the technology sector for a decade. He will add depth to the team, focusing on areas including corporate governance and long-term business initiatives.”

Prior to joining ByteDance earlier this year, Chew was an executive at Xiaomi, with stints at DST and Goldman Sachs earlier in his career.

 



from Social – TechCrunch https://ift.tt/3u6vbfS
via IFTTT

Computer vision inches towards ‘common sense’ with Facebook’s latest research

Machine learning is capable of doing all sorts of things as long as you have the data to teach it how. That’s not always easy, and researchers are always looking for a way to add a bit of “common sense” to AI so you don’t have to show it 500 pictures of a cat before it gets it. Facebook’s newest research takes a big step towards reducing the data bottleneck.

The company’s formidable AI research division has been working on how to advance and scale things like advanced computer vision algorithms for years now, and has made steady progress, generally shared with the rest of the research community. One interesting development Facebook has pursued in particular is what’s called “semi-supervised learning.”

Generally when you think of training an AI, you think of something like the aforementioned 500 pictures of cats — images that have been selected and labeled (which can mean outlining the cat, putting a box around the cat, or just saying there’s a cat in there somewhere) so that the machine learning system can put together an algorithm to automate the process of cat recognition. Naturally if you want to do dogs or horses, you need 500 dog pictures, 500 horse pictures, etc — it scales linearly, which is a word you never want to see in tech.

Semi-supervised learning, related to “unsupervised” learning, involves figuring out important parts of a dataset without any labeled data at all. It doesn’t just go wild, there’s still structure; for instance, imagine you give the system a thousand sentences to study, then showed it ten more that have several of the words missing. The system could probably do a decent job filling in the blanks just based on what it’s seen in the previous thousand. But that’s not so easy to do with images and video — they aren’t as straightforward or predictable.

But Facebook researchers have shown that while it may not be easy, it’s possible and in fact very effective. The DINO system (which stands rather unconvincingly for “DIstillation of knowledge with NO labels”) is capable of learning to find objects of interest in videos of people, animals, and objects quite well without any labeled data whatsoever.

Animation showing four videos and the AI interpretation of the objects in them.

Image Credits: Facebook

It does this by considering the video not as a sequence of images to be analyzed one by one in order, but as an complex, interrelated set,like the difference between “a series of words” and “a sentence.” By attending to the middle and the end of the video as well as the beginning, the agent can get a sense of things like “an object with this general shape goes from left to right.” That information feeds into other knowledge, like when an object on the right overlaps with the first one, the system knows they’re not the same thing, just touching in those frames. And that knowledge in turn can be applied to other situations. In other words, it develops a basic sense of visual meaning, and does so with remarkably little training on new objects.

This results in a computer vision system that’s not only effective — it performs well compared with traditionally trained systems — but more relatable and explainable. For instance, while an AI that has been trained with 500 dog pictures and 500 cat pictures will recognize both, it won’t really have any idea that they’re similar in any way. But DINO — although it couldn’t be specific — gets that they’re similar visually to one another, more so anyway than they are to cars, and that metadata and context is visible in its memory. Dogs and cats are “closer” in its sort of digital cognitive space than dogs and mountains. You can see those concepts as little blobs here — see how those of a type stick together:

Animated diagram showing how concepts in the machine learning model stay close together.

Image Credits: Facebook

This has its own benefits, of a technical sort we won’t get into here. If you’re curious, there’s more detail in the papers linked in Facebook’s blog post.

There’s also an adjacent research project, a training method called PAWS, which further reduces the need for labeled data. PAWS combines some of the ideas of semi-supervised learning with the more traditional supervised method, essentially giving the training a boost by letting it learn from both the labeled and unlabeled data.

Facebook of course needs good and fast image analysis for its many user-facing (and secret) image-related products, but these general advances to the computer vision world will no doubt be welcomed by the developer community for other purposes.



from Social – TechCrunch https://ift.tt/3vxpTdA
via IFTTT

Rapchat tunes into $2.3M as its music-making app hits 7M users

YouTube, Snapchat, Twitter, TikTok and Facebook’s Instagram have upended the film and TV industries, with a new wave of cinematographers, directors and actors leveraging innovations in technology to create new work and connect directly with billions of consumers to see it. Today, a startup is announcing some funding as it looks to make a similar impact in the world of music.

Rapchat, an app that lets people create music tracks — raps, as its names suggest, or something else — using a platform that crowdsources beats and lets people put vocals on top of them, has raised $2.3 million.

Co-led by Sony Music Entertainment and NYC VC firm Adjacent, this is an extension to Rapchat’s seed round of $1.7 million back in 2018, and CEO and co-founder Seth Miller tells me it’s coming as the startup is getting ready for a bigger Series A.

With no connection to Snapchat — not now at least, except that founders Seth Miller and Pat Gibson did think it was a funny pun at the time that they were first conceiving of the company as a side hustle while still in university in 2015 — Rapchat has already gone quite some way in scaling.

The company today has some 7 million registered users, and at the moment some 250,000 songs are being created around a catalog of about 100,000 beats by 500,000 active users on the platform each month. Engagement is hovering right now at 35 minutes per day on average, a mix not just of people making tunes, but through the beginnings of a social graph: people coming onto the app to discover and share those tracks.

Rapchat plans to use the funding to continue expanding the scope of what you can create on its platform, including growing the prize pools for Rapchat’s ‘Challenges’ competition series; expand to have more artists, producers, and industry executives on the platform for mentoring; and to extend that platform’s reach to integrate more deeply with the likes of TikTok, Snapchat, Spotify and Apple Music — platforms where creators are already making a lot of content, and where music is figuring strongly in that effort.

Rapchat’s growth not only speaks to how the startup has pulled off its ambition to make it easier to make music, but it also speaks to an appetite, an itch, in the creator economy: there is a big wide world of music-making out there, and more want to see if they can strike the right note.

Rapchat is definitely not the only, nor the first, company to think of how to address music creators within the bigger creator economy.

Another app called Voisey had conceived of a similar idea but focused primarily on letting people create and record shorter clips rather than full music tracks before sharing them to other platforms. It has not quite come a household name, but it did have some small success in bringing attention to new artists, and interestingly, it was quietly acquired by Snap last year (and for now Snap’s kept Voisey’s app up).

TikTok’s parent ByteDance has also made an acquisition of another music creation app, Jukedeck. As with Snap’s acquisition, so far we’re not fully clear on how and where that acquisition is going, but we’ve heard through the grapevine that TikTok is working on a new music service that sounds like it might let more content get plugged into TikTok’s music layer, so perhaps watch this space.

And in perhaps the most trend-endorsing act of all, Rapchat has been cloned — by Facebook, no less. NPE, the social networking behemoth’s in-house skunkworks team, in February rolled out BARS (all caps! stand out!) — which is, yes — an app on which you can create your own rap music.

Miller, at least for now, is about as laid back as you could be, considering all of the above, confident that at least for now, he is very happy with the engagement Rapchat is seeing, including around tests it has been running around offering new premium features — the app is free to use right now, but it has plans to offer creators more production tools and better ways of sharing their work and helping build a business out of it. Key to that will be never demanding licensing fees on music: creators keep the royalties, with Rapchat’s value lying in helping them make and track how that music gets used with the metadata that it holds on those tracks.

Some of the low-key approach might well come from the fact that Rapchat and its founders are somewhat outside of the startup fray. The idea for the app first came up in 2013, Miller said, when he and Gibson were students at Ohio University in Columbus.

“We were coming of age when everyone in college was using apps like Snapchat and Instagram,” he said. “We loved them for video, but saw there was nothing like them for creating music. So we pitched the idea during a Startup Weekend competition: snapping like Snapchat but for rap. Someone said, ‘Rapchat’ and we liked it.”

They went full-time on the idea in 2015 when they got into 500 Startups with the app, but even so it’s taken them years to build up the business, get attention from investors and raise money. Why? Partly because music is hard, and frankly the main game in town for years has been streaming services, rather than creation services.

Miller and Gibson persisted: “I knew that this market was huge. It just made so much sense to me,” he said. “The advent of the mobile devices the moment that apps like Instagram, VSCO and Snapchat have turned people into photographers and video makers, and Substack is turning people into writers.” And now Rapchat wants to tap the world for rappers.

“Rapchat has created a music studio that fits into your pocket,” said Nico Wittenborn, lead Investor at venture capital firm Adjacent, in a statement. “It decreases the friction of creativity by allowing anyone, anywhere in the world to record and publish music straight from their phones. This mobile-enabled democratization of technology is what Adjacent is all about, and I am super excited to support the team in building out this next-level music platform.”



from Social – TechCrunch https://ift.tt/3u8R12j
via IFTTT

Instagram Live takes on Clubhouse with options to mute and turn off the video

In addition to Facebook’s Clubhouse competitor built within Messenger Rooms and its experiments with a Clubhouse-like Q&A platform on the web, the company is now leveraging yet another of its largest products to take on the Clubhouse threat: Instagram Live. Today, Instagram announced it’s adding new features that will allow users to mute their microphones and even turn their video off while using Instagram Live.

Instagram explains these new features will give hosts more flexibility during their livestream experiences, as they can decrease the pressure to look or sound a certain way while broadcasting live. While that may be true, the reality is that Facebook is simply taking another page from Clubhouse’s playbook by enabling a “video off” experience that encourages more serendipitous conversations.

When people don’t have to worry about how they look, they’ll often be more amenable to jumping into a voice chat. In addition, being audio-only allows creators to engage with their community while multitasking — perhaps they’re doing chores or moving around, and can’t sit and stare right at the camera. To date, this has been one of the advantages about using Clubhouse versus live video chat. You could participate in Clubhouse’s voice chat rooms without always having to give the conversation your full attention or worrying about background noise.

For the time being, hosts will not be able to turn on or off the video or mute others in the livestream, but Instagram tells us it’s working on offering more of these types of capabilities to the broadcaster, and expects to roll them out soon.

Instagram notes it tested the new features publicly earlier this week during an Instagram Live between Facebook CEO Mark Zuckerberg and Head of Instagram Adam Mosseri.

This isn’t the first feature Instagram has added in recent weeks to lure the creator community to its platform instead of Clubhouse or other competitors. In March, Instagram rolled out the option for creators to host Live Rooms that allow up to four people to broadcast at the same time. The Rooms were meant to appeal to creators who wanted to host live talk shows, expanded Q&As, and more — all experiences that are often found on Clubhouse. It also added the ability for fans to buy badges to support the hosts, to cater to the needs of professional creators looking to monetize their reach.

Although Instagram parent company Facebook already has a more direct Clubhouse clone in development with Live Audio Rooms on Facebook and Messenger, the company said it doesn’t expect it to launch into testing until this summer. And it will first be available to Groups and public figures, not the broader public.

Instagram Live’s new features, meanwhile, are rolling out to Instagram’s global audience on both iOS and Android starting today.



from Social – TechCrunch https://ift.tt/3xt3LTH
via IFTTT

EU adopts rules on one-hour takedowns for terrorist content

The European Parliament approved a new law on terrorist content takedowns yesterday, paving the way for one-hour removals to become the legal standard across the EU.

The regulation “addressing the dissemination of terrorist content online” will come into force shortly after publication in the EU’s Official Journal — and start applying 12 months after that.

The incoming regime means providers serving users in the region must act on terrorist content removal notices from Member State authorities within one hour of receipt, or else provide an explanation why they have been unable to do so.

There are exceptions for educational, research, artistic and journalistic work — with lawmakers aiming to target terrorism propaganda being spread on online platforms like social media sites.

The types of content they want speedily removed under this regime includes material that incites, solicits or contributes to terrorist offences; provides instructions for such offences; or solicits people to participate in a terrorist group.

Material posted online that provides guidance on how to make and use explosives, firearms or other weapons for terrorist purposes is also in scope.

However concerns have been raised over the impact on online freedom of expression — including if platforms use content filters to shrink their risk, given the tight turnaround times required for removals.

The law does not put a general obligation on platforms to monitor or filter content but it does push service providers to prevent the spread of proscribed content — saying they must take steps to prevent propagation.

It is left up to service providers how exactly they do that, and while there’s no legal obligation to use automated tools it seems likely filters will be what larger providers reach for, with the risk of unjustified, speech chilling takedowns fast-following. 

Another concern is how exactly terrorist content is being defined under the law — with civil rights groups warning that authoritarian governments within Europe might seek to use it to go after critics based elsewhere in the region.

The law does include transparency obligations — meaning providers must publicly report information about content identification and takedown actions annually.

On the sanctions side, Member States are responsible for adopting rules on penalties but the regulation sets a top level of fines for repeatedly failing to comply with provisions at up to 4% of global annual turnover.

EU lawmakers proposed the new rules back in 2018  when concern was riding high over the spread of ISIS content online.

Platforms were pressed to abide by an informal one-hour takedown rule in March of the same year. But within months the Commission came with a more expansive proposal for a regulation aimed at “preventing the dissemination of terrorist content online”.

Negotiations over the proposal have seen MEPs and Member States (via the Council) tweaking provisions — with the former, for example, pushing for a provision that requires competent authority to contact companies that have never received a removal order a little in advance of issuing the first order to remove content — to provide them with information on procedures and deadlines — so they’re not caught entirely on the hop.

The impact on smaller content providers has continued to be a concern for critics, though.

The Council adopted its final position in March. The approval by the Parliament yesterday concludes the co-legislative process.

Commenting in a statement, MEP Patryk JAKI, the rapporteur for the legislation, said: “Terrorists recruit, share propaganda and coordinate attacks on the internet. Today we have established effective mechanisms allowing member states to remove terrorist content within a maximum of one hour all around the European Union. I strongly believe that what we achieved is a good outcome, which balances security and freedom of speech and expression on the internet, protects legal content and access to information for every citizen in the EU, while fighting terrorism through cooperation and trust between states.”



from Social – TechCrunch https://ift.tt/3xA6wmc
via IFTTT

Facebook hides posts calling for PM Modi’s resignation in India

Facebook has temporarily hidden all posts with hashtag “ResignModi” in India, days after the U.S. social juggernaut — along with Twitter — complied with an order from New Delhi to censor some posts critical of Indian government’s handling of the coronavirus pandemic.

On its website, Facebook said it has hidden posts with “ResignModi” hashtag because some of those violated its community standards. (A search for “ResignModi” is currently returning some results for users in the U.S.) It’s unclear at this time if Facebook was ordered to take this call or if it did so at its own will.

Tweets with “#ResignModi”, at the time of publication, were visible in India. With over 450 million WhatsApp users and nearly 400 million Facebook users, India is the largest market for the social company by the size of userbase.

Covid cases have surged in the South Asian nation in recent days, prompting many citizens to air their frustration toward the government on social channels as they struggle to find an empty bed, oxygen supplies and medicines in hospitals.

Image Credits: Screengrab by TechCrunch

Facebook, which didn’t respond to a request for comment over censorship of posts in India over the weekend, didn’t immediately respond to a request for comment Wednesday.

This is a developing story. More to follow…



from Social – TechCrunch https://ift.tt/3ue3IJv
via IFTTT

Telegram to add group video calls next month

Group video calls will be coming to Telegram’s messaging platform next month with what’s being touted as a fully featured implementation, including support for web-based videoconferencing.

Founder Pavel Durov made the announcement via a (text) message posted to his official Telegram channel today where he wrote “we will be adding a video dimension to our voice chats in May, making Telegram a powerful platform for group video calls”.

“Screen sharing, encryption, noise-cancelling, desktop and tablet support — everything you can expect from a modern video conferencing tool, but with Telegram-level UI, speed and encryption. Stay tuned!” he added, using the sorts of phrases you’d expect from an enterprise software maker.

Telegram often taunts rivals over their tardiness to add new features but on video calls it has been a laggard, only adding the ability to make one-on-one video calls last August — rather than prioritizing a launch of group video calls, as it had suggested it would a few months earlier.

In an April 2020 blog post, to mark passing 400M users, it wrote that the global lockdown had “highlighted the need for a trusted video communication tool” — going on to dub video calls in 2020 “much like messaging in 2013”.

However it also emphasized the importance of security for group video calling — and that’s perhaps what’s caused the delay.

(Another possibility is the operational distraction of needing to raise a large chunk of debt financing to keep funding development: Last month Telegram announced it had raised over $1BN by selling bonds — its earlier plan to monetize via a blockchain platform having hit the buffers in 2020.)

In the event, rather than rolling out group video calls towards the latter end of 2020 it’s going to be doing so almost half way through 2021 — which has left videoconferencing platforms like Zoom to keep cleaning up during the pandemic-fuelled remote work and play boom (even as ‘Zoom fatigue’ has been added to our lexicon).

How secure Telegram’s implementation of group video calls will be, though, is an open question.

Durov’s post mades repeat mention of “encryption” — perhaps to make a subtle dig at Zoom’s own messy security claims history — but doesn’t specify whether it will use end-to-end encryption (we’ve asked).

Meanwhile Zoom does now offer e2e — and also has designs on becoming a platform in its own right, with apps and a marketplace, so there are a number of shifts in the comms landscape that could see the videoconferencing giant making deeper incursions into Telegram’s social messaging territory.

The one-to-one video calls Telegram launched last year were rolled out with its own e2e encryption — so presumably it will be replicating that approach for group calls.

However the MTProto encryption Telegram uses is custom-designed — and there’s been plenty of debate among cryptography experts over the soundness of its approach. So even if group calls are e2e encrypted there will be scrutiny over exactly how Telegram is doing it.

Also today, Durov touted two recently launched web versions of Telegram (not the first such versions by a long chalk, though) — adding that it’s currently testing “a functional version of web-based video calls internally, which will be added soon”.

He said the Webk and Webz versions of the web app are “by far the most cross-platform versions of Telegram we shipped so far”, and noting that no downloads or installs are required to access your chats via the browser.

“This is particularly good for corporate environments where installing native apps is now always allowed, but also good for users who like the instant nature of web sites,” he added, with another little nod toward enterprise users.



from Social – TechCrunch https://ift.tt/3gR3VOV
via IFTTT

Social networking app for women, Peanut, adds live audio rooms

Mobile social networking app for women, Peanut, is today becoming the latest tech company to integrate audio into its product following the success of Clubhouse. Peanut, which began with a focus on motherhood, has expanded over the years to support women through all life stages, including pregnancy, marriage and even menopause. It sees its voice chat feature, which it’s calling “Pods,” as a way women on its app can make better connections in a more supportive, safer environment than other platforms may provide.

The pandemic, of course, likely drove some of the interest in audio-based social networking, as people who had been stuck at home found it helped to fill the gap that in-person networking and social events once did. However, voice chat social networking leader Clubhouse has since seen its model turned into what’s now just a feature for companies like Facebook, Twitter, Reddit, LinkedIn, Discord and others to adopt.

Like many of the Clubhouse clones to date, Peanut’s Pods offer the basics, including a muted audience of listeners who virtually “raise their hand” to speak, emoji reactions and hosts who can moderate the conversations and invite people to speak, among other things. The company, for now, is doing its own in-house moderation on the audio pods, to ensure the conversations don’t violate the company’s terms. In time, it plans to scale to include other moderators. (The company pays over two dozen moderators to help it manage the rest of its app, but the team had not yet been trained on audio, as of just a few days ago. They have now been given the training, we understand.)

Though there are similarities with Clubhouse in its design, what Peanut believes will differentiate its audio experience from the rest of the pack is where these conversations are taking place — on a network designed for women built with safety and trust in mind. It’s also a network where chasing clout is not the reason people participate.

Traditional social networks are often based on how many likes you have, how many followers you have, or if you’re verified with a blue check, explains Peanut founder CEO Michelle Kennedy.

“It’s kind of all based around status and popularity,” she says. “What we’ve only ever seen on Peanut is this ‘economy of care,’ where women are really supportive of one another. It’s really never been about, ‘I’ve got X number of followers.’ We don’t even have that concept. It’s always been about: ‘I need support; I have this question; I’m lonely or looking for a friend;’ or whatever it might be,” Kennedy adds.

In Peanut Pods, the company says it will continue to enforce the safety standards that make women feel comfortable with social networking. This focus in particular could attract some of the women, and particularly women of color, who have been targeted with harassment on other voice-based networking platforms.

“The one thing I would say is we’re a community, and we have standards,” notes Kennedy. “When you have standards and you let everyone know what those standards are, it’s very clear. You’re allowed an opinion but what you’re not allowed to do are listed here…Here are the things we expect of you as a user and we’ll reward you if you do it and if you don’t, we’re going to ask you to leave,” she says.

Freedom of speech is not what Peanut’s about, she adds.

“We have standards and we ask you to adhere to them,” says Kennedy.

In time, Peanut envisions using the audio feature to help connect women with people who have specific expertise, like lactation consultants for new moms or fertility doctors, for example. But these will not be positioned as lectures where listeners are held hostage as a speaker drones on and on. In fact, Peanut’s design does away with the “stage” concept from Clubhouse to give everyone equal status — whether they’re speaking or not.

In the app, users will be able to find interesting chats based on what topics they’re already following — and, importantly, they can avoid being shown other topics by muting them.

The Pods feature is rolling out to Peanut’s app starting today, where it will reach the company’s now 2 million-plus users. It will be free to use, like all of Peanut, though the company plans to eventually launch a freemium model with some paid products further down the road.

Update, 4/27/21, 3:40 PM ET: Article updated to note moderators have now received Pods training. 



from Social – TechCrunch https://ift.tt/3xtiwpu
via IFTTT

TikTok to open a ‘Transparency’ Center in Europe to take content and security questions

TikTok will open a center in Europe where outside experts will be shown information on how it approaches content moderation and recommendation, as well as platform security and user privacy, it announced today.

The European Transparency and Accountability Centre (TAC) follows the opening of a U.S. center last year — and is similarly being billed as part of its “commitment to transparency”.

Soon after announcing its U.S. TAC, TikTok also created a content advisory council in the market — and went on to replicate the advisory body structure in Europe this March, with a different mix of experts.

It’s now fully replicating the U.S. approach with a dedicated European TAC.

To-date, TikTok said more than 70 experts and policymakers have taken part in a virtual U.S. tour, where they’ve been able to learn operational details and pose questions about its safety and security practices.

The short-form video social media site has faced growing scrutiny over its content policies and ownership structure in recent years, as its popularity has surged.

Concerns in the U.S. have largely centered on the risk of censorship and the security of user data, given the platform is owned by a Chinese tech giant and subject to Internet data laws defined by the Chinese Communist Party.

While, in Europe, lawmakers, regulators and civil society have been raising a broader mix of concerns — including around issues of child safety and data privacy.

In one notable development earlier this year, the Italian data protection regulator made an emergency intervention after the death of a local girl who had reportedly been taking part in a content challenge on the platform. TikTok agreed to recheck the age of all users on its platform in Italy as a result.

TikTok said the European TAC will start operating virtually, owing to the ongoing COVID-19 pandemic. But the plan is to open a physical center in Ireland — where it bases its regional HQ — in 2022.

EU lawmakers have recently proposed a swathe of updates to digital legislation that look set to dial up emphasis on the accountability of AI systems — including content recommendation engines.

A draft AI regulation presented by the Commission last week also proposes an outright ban on subliminal uses of AI technology to manipulate people’s behavior in a way that could be harmful to them or others. So content recommender engines that, for example, nudge users into harming themselves by suggestively promoting pro-suicide content or risky challenges may fall under the prohibition. (The draft law suggests fines of up to 6% of global annual turnover for breaching prohibitions.)

It’s certainly interesting to note TikTok also specifies that its European TAC will offer detailed insight into its recommendation technology.

“The Centre will provide an opportunity for experts, academics and policymakers to see first-hand the work TikTok teams put into making the platform a positive and secure experience for the TikTok community,” the company writes in a press release, adding that visiting experts will also get insights into how it uses technology “to keep TikTok’s community safe”; how trained content review teams make decisions about content based on its Community Guidelines; and “the way human reviewers supplement moderation efforts using technology to help catch potential violations of our policies”.

Another component of the EU’s draft AI regulation sets a requirement for human oversight of high risk applications of artificial intelligence. Although it’s not clear whether a social media platform would fall under that specific obligation, given the current set of categories in the draft regulation.

However the AI regulation is just one piece of the Commission’s platform-focused rule-making.

Late last year it also proposed broader updates to rules for digital services, under the DSA and DMA, which will place due diligence obligations on platforms — and also require larger platforms to explain any algorithmic rankings and hierarchies they generate. And TikTok is very likely to fall under that requirement.

The UK — which is now outside the bloc, post-Brexit — is also working on its own Online Safety regulation, due to present this year. So, in the coming years, there will be multiple content-focused regulatory regimes for platforms like TikTok to comply with in Europe. And opening algorithms to outside experts may be hard legal requirement, not soft PR.

Commenting on the launch of its European TAC in a statement, Cormac Keenan, TikTok’s head of trust and safety, said: With more than 100 million users across Europe, we recognise our responsibility to gain the trust of our community and the broader public. Our Transparency and Accountability Centre is the next step in our journey to help people better understand the teams, processes, and technology we have to help keep TikTok a place for joy, creativity, and fun. We know there’s lots more to do and we’re excited about proactively addressing the challenges that lie ahead. I’m looking forward to welcoming experts from around Europe and hearing their candid feedback on ways we can further improve our systems.”

 



from Social – TechCrunch https://ift.tt/32TexEK
via IFTTT

The next tech hearing targets social media algorithms — and YouTube, for once

Another week, another big tech hearing in Congress. With a flurry of antitrust reform bills on the way, Democratic lawmakers are again bringing in some of the world’s most powerful tech companies for questioning.

In the next hearing, scheduled for Tuesday, April 27 at 10 AM ET, the Senate Judiciary’s subcommittee on privacy and technology will zero in on concerns about algorithmic amplification. Specifically, the hearing will explore how algorithms amplify dangerous content and shape user behavior on social platforms.

The subcommittee’s chair Sen. Chris Coons previously indicated that he would bring in tech CEOs, but Tuesday’s hearing will instead feature testimony from policy leads at Facebook, Twitter and YouTube.

The hearing might prove a unique opportunity to hold YouTube’s feet to the fire. In spite of being one of the biggest social networks in the world — one without much transparency about its regular failures to control extremism and misinformation — YouTube seldom winds up under the microscope with Congress. The company will be represented by Alexandra Veitch, YouTube’s regional director of public policy.

In past big tech hearings, Google CEO Sundar Pichai has generally appeared on behalf of YouTube’s parent company while YouTube’s chief executive Susan Wojcicki inexplicably escapes scrutiny. Google is a massive entity and concerns specific to YouTube and its policies generally get lost in the mix, with lawmakers usually going after Pichai for concerns around Google’s search and ads businesses.

In a stylistic repeat of last week’s adversarial app store hearing, which featured Apple as well as some of its critics, misinformation researcher Dr. Joan Donovan and ex-Googler and frequent big tech critic Tristan Harris will also testify Tuesday. That tension can create deeper questioning, providing outside expertise that can fill in some lapses in lawmakers’ technical knowledge.

Policy leads at these companies might not make the same flashy headlines, but given their intimate knowledge of the content choices these companies make every day, they do provide an opportunity for more substance. Tech CEOs like Mark Zuckerberg and Jack Dorsey have been dragged in to so many hearings at this point that they begin to run together, and the top executives generally reveal very little while sometimes playing dumb about the day-to-day decision making on their platforms. The subcommittee’s ranking member Ben Sasse (R-NE) emphasized that point, stating that the hearing would be a learning opportunity and not a “show hearing.”

Democrats have been sounding the alarm on algorithms for some time. While Republicans spent the latter half of the Trump administration hounding tech companies about posts they remove, Democrats instead focused on the violent content, extremism and sometimes deadly misinformation that gets left up and even boosted by the secretive algorithms tech companies rarely shed light on.

We haven’t seen much in the way of algorithmic transparency, but that could change. One narrowly targeted Section 230 reform bill in the House would strip that law’s protections from large companies when their algorithms amplify extremism or violate civil rights.

Twitter CEO Jack Dorsey has also hinted that a different approach might be on the horizon, suggesting that users could hand-pick their preferred algorithms in the future, possibly even selecting them from a kind of third-party marketplace. Needless to say, Facebook didn’t indicate any plans to give its own users more algorithmic control.

With any major changes to the way platforms decide who sees what likely a long ways off, expect to see lawmakers try to pry open some black boxes on Tuesday.



from Social – TechCrunch https://ift.tt/2QUDblv
via IFTTT

Dating app S’More bets on celebrity content with S’More TV

Dating startup S’More has launched a new feature called S’More TV, which CEO Adam Cohen-Aslatei said could give users new ways to talk to each other. It also bringing the startup closer to his goal of becoming not just a dating app, but “a lifestyle brand.”

The videos on S’More TV may be familiar to people who follow the company on Instagram, where Cohen-Aslatei has hosted dating-related interviews with celebrities like WWE stars Chelsea Green and Leah Van Dale, models Nichole Holms and Sarah Miller and Mary Fitzgerald of “Selling Sunset.” Now, however, these videos are getting a home in the S’More app itself.

Cohen-Aslatei said this wasn’t the initial plan when he started filming videos for S’More Live, but eventually he and his team decided to do more with their “nearly 50 hours of exclusive celebrity content,” seeing them as “the first organic way to have content not related to a dating profile” in the app. He suggested that this isn’t just giving users another reason to open the app, but also a crucial conversation starter.

After all, anyone who’s had to start a conversation in a dating app will probably remember moments when you’ve struggled to come up with something better than “Hey” or “How’s it going?” — and S’More presents unique challenges on that front, since the app blurs all user photos until you’ve had some real interaction. So Cohen-Aslatei said these videos can spur a more natural conversation, allowing users to “really talk about topics that they care about.”

S'More TV

Image Credits: S’More

The S’More app now includes prompts directing users to watch S’More TV, where anyone can watch and comment on the videos. If you post a particularly scintillating comment, that may help to attract potential matches, and it will give them an easy starting point for the conversation.

In fact, Cohen-Aslatei claimed that since releasing S’More TV in beta testing, the app has seen Day 1 retention (the number of users who open the app one day after installation) increase by 15% to the “mid 60s,” while time in the app has doubled.

He also said S’More Live is quickly approaching its 100th episode, but you won’t see all of those episodes the app right away. Instead, the company will be adding 15 new videos to the S’More app each month.

In addition, S’More has launched a new feature called cover photos. This is basically the one exception to the blurred photos rule (although it sounds like even this photo is subtly warped), allowing users to showcase a single image of their friends, their life or their personality even before they start a conversation.



from Social – TechCrunch https://ift.tt/32NFSYX
via IFTTT

Facebook introduces a new miniplayer that streams Spotify within the Facebook app

Facebook announced last week an expanded partnership with streaming music service Spotify that would bring a new way to listen to music or podcasts directly within Facebook’s app, which it called Project Boombox. Today, the companies are rolling out this integration via a new “miniplayer” experience that will allow Facebook users to stream from Spotify through the Facebook app on iOS or Android. The feature will be available to both free Spotify users and Premium subscribers.

The miniplayer itself is an extension of the social sharing option already supported within Spotify’s app. Now, when Spotify users are listening to content they want to share to Facebook, they’ll be able to tap the existing “Share” menu (the three dot-menu at the upper right of the screen) and then tap either “Facebook” or “Facebook News Feed.”

When a user posts an individual track or podcast episode to Facebook through this sharing feature, the post will now display in a new miniplayer that allows other people who come across their post to also play the content as they continue to scroll, or reshare it. (Cue MySpace vibes!)

Spotify’s paid subscribers will be able to access full playback, the company says. Free users, meanwhile, will be able to hear the full shared track, not a clip . But afterwards, they’ll continue to listen to ad-supported content on Shuffle mode, just as they would in Spotify’s own app.

One important thing to note here about all this works is that the integration allows the music or podcast content to actually play from within the Spotify app. When a user presses play on the miniplayer, an app switch takes place so the user can log into Spotify. The miniplayer activates and controls the launch and playback in the Spotify app — which is how the playback is able continue even as the user scrolls on Facebook or if they minimize the Facebook app altogether.

This setup means users will need to have the Spotify mobile app installed on their phone and a Spotify account for the miniplayer to work. For first-time Spotify users, they’ll have to sign up for a free account in order to listen to the music shared via the miniplayer.

Spotify notes that it’s not possible to sign up for a paid account through the mini-player experience itself, so there’s no revenue share with Facebook on new subscriptions. (Users have to download the Spotify app and sign up for Paid accounts from there if they want to upgrade.)

The partnership allows Spotify to leverage Facebook’s reach to gain distribution and to drive both sign-ups and repeat usage of its app just as the Covid bump to subscriber growth may be wearing off. However, it’s still responsible for the royalties paid on streams, just as it was before, the company told TechCrunch, because its app is the one actually doing the streaming. It’s also fully in charge of the music catalog and audio ads that play alongside the content.

For Facebook, this deal means it now has a valuable tool to keep users spending time on its site — a metric that has been declining over the years, reports have indicated.

Spotify and Facebook have a long history of working together on music efforts. Facebook back in 2011 had been planning an update that would allow music subscription users to engage with music directly on Facebook, much like this. But those plans were later dialed back, possibly over music rights’ or technical issues. Spotify had also been one of the first media partners on Facebook’s ticker, which would show you in real-time what friends were up to on Facebook and other services. And Spotify had once offered Facebook Login as the default for its mobile app. Today, as it has for years, Spotify users on the desktop can see what their Facebook friends are streaming on its app, thanks to social networking integrations.

The timing for this renewed and extended partnership is interesting. Now, both Facebook and Spotify have a mutual enemy with Apple, whose privacy-focused changes are impacting Facebook’s ad business and whose investments in Apple Music and Podcasts are a threat to Spotify. As Facebook’s own music efforts in more recent years have shifted towards partnership efforts — like music video integrations enabled by music label agreements — it makes sense that it would turn to a partner like Spotify to power a new streaming feature that supports Facebook’s broader efforts around monetizable tools and services aimed at the creator economy.

The miniplayer feature had been tested in non-U.S. markets, Mexico and Thailand, ahead of its broader global launch today.

In addition to the U.S., the integration is fully rolling out to users in Argentina, Australia, Bolivia, Brazil, Canada, Chile, Colombia, Costa Rica, Dominican Republic, Ecuador, El Salvador, Guatemala, Honduras, Indonesia, Israel, Japan, Malaysia, Mexico, New Zealand, Nicaragua, Panama, Paraguay, Peru, South Africa, Thailand, and Uruguay.



from Social – TechCrunch https://ift.tt/3nlxOYJ
via IFTTT

India orders Twitter to take down tweets critical of its coronavirus handling

Twitter has taken down dozens of tweets in India, some of which were critical of New Delhi’s handling of the coronavirus, to comply with an emergency order from the Indian government.

New Delhi made an emergency order to Twitter to censor over 50 tweets in the country, Twitter disclosed on Lumen database. The social network has complied with the request, and withheld those tweets from users in India.

TechCrunch has learned that Twitter is not the only platform affected by the new order.

India, which has also previously ordered Twitter to take down some tweets and accounts critical of its policies and threatened jail time to employees in the event of non-compliance, comes as the country reports a record of over 330,000 new covid cases a day, the worst by any country. News reports suggest that even this number is underreported.

Amid a collapse of the nation’s health infrastructure, Twitter has become a rare beam of hope as people crowdsource data and help one another find medicines and oxygen cylinders.

A copy of one of Indian government’s orders disclosed by Twitter. (Lumen Database)

Medianama, which first reported on New Delhi’s new order, said among those whose tweets have been censored in India include Revanth Reddy (a Member of Parliament), Moloy Ghatak (a minister in West Bengal), Vineet Kumar Singh (actor) filmmakers Vinod Kapri and Avinash Das.

In a statement, a Twitter spokesperson said, “When we receive a valid legal request, we review it under both the Twitter Rules and local law. If the content violates Twitter’s Rules, the content will be removed from the service. If it is determined to be illegal in a particular jurisdiction, but not in violation of the Twitter Rules, we may withhold access to the content in India only. In all cases, we notify the account holder directly so they’re aware that we’ve received a legal order pertaining to the account.”

“We notify the user(s) by sending a message to the email address associated with the account(s), if available. Read more about our Legal request FAQs.  The legal requests that we receive are detailed in the bianual Twitter Transparency Report, and requests to withhold content are published on Lumen.”



from Social – TechCrunch https://ift.tt/32Lf5fC
via IFTTT

Filing: Snap paid $124M for Fit Analytics as it gears up for a bigger e-commerce push

Earlier this year we reported on how Snap had acquired Berlin-based Fit Analytics, an AI-based fitting technology startup, as part of a wider push into e-commerce services, specifically to gain technology that can help prospective online shoppers get a better sense of how a particular item or size would fit them. A 10-Q filing from Snap today has now put a price tag on that deal.

Snap paid a total of $124.4 million, covering technology, IP, customer relationships and payouts to the team. The filing also noted that Snap spent a total of $204.5 million on acquisitions in 2020, but did not break them out.

The news comes ahead of Snap — whose flagship app Snapchat now has 280 million daily active users — preparing for its Snap Partner Conference in May. Sources say the company plans to announce, among other news, deeper commerce features for Snapchat — specifically tools to make it easier for Snapchat users to interact with and buy items that appear in the app, either in ads or more organically in content shared by other users.

While the exact details of those commerce tools, and the timing of when they might come online, are not yet known, Snap has hardly kept its interest in commerce a secret.

Snap has been hiring for roles to support its commerce efforts. Currently it’s advertising for a variety of engineering, marketing and product roles in commerce, to, in the words of one of the listings, for a Product Manager, “develop and launch shopping experiences and services that make shopping fun for Snapchatters and drive results for brands.” The listings also include a role specifically to work on Snapchat-based e-commerce efforts for direct-to-consumer (D2C) businesses.

And it has been making other recent acquisitions in addition to Fit Analytics that also line up with that.

They have included Screenshop, an app that describes itself as “the first AI-back style lens,” which can identify shoppable items in photos and then build a custom catalog of similar products that you can buy (akin to “shop the look” features that you will have come across in fashion media). And it’s also acquired Ariel AI, which has built technology to quickly render people in 3D, technology that can be used in a diverse set of applications, from games to virtual try-ons of clothing, makeup or accessories.

Snap confirmed the Ariel acquisition to CNBC in January. And while Screenshop deal was first reported earlier this month by The Information, Snap has declined to comment on it, although we have found people who worked at the startup now working at Snap.

Both acquisitions closed in 2020, according to reports, meaning that they came out of that year’s $204.5 million acquisition run. (Snap also noted a smaller acquisition, for $7.6 million, in the most recent quarter, but it did not disclose any further details.)

Even before all this, Snap had been making smaller efforts and tests in commerce going back years, although none of them have tipped into mainstream efforts.

Among them, in 2018 it launched a Snap Store — but that so far has not progressed beyond selling merchandise based on Bitmoji characters. And work on a Gucci shoe campaign last year, where Snapchat users could try shoes on in AR and then buy them, was seen by some as its big step into commerce — “we’ve moved from pure entertainment and expanded the use-case. And so with brands, it’s a really exciting time, especially in fashion and beauty. The Snapchat camera is connecting brands to their audiences in new ways,” a Snapchat AR executive said at the time — but that also didn’t develop into much beyond a one-off effort.

But with the pandemic leading to a surge of shopping online, and technology continuing to improve, the iron may finally be hot here.

As we said around the Fit Analytics acquisition, the idea of diversifying Snapchat’s revenue streams by building in more commerce experiences makes a lot of sense.

It gives the company another revenue stream at a time when Apple is introducing changes that might well affect how advertising can run and be monetized in the future. (The company most recently posted average revenues per user of $2.74, a figure Wall Street will be hoping will grow, not shrink.) It also plays into the demographics that Snapchat targets, where younger consumers are using social media apps to discover, share and shop for goods.

And specifically in the case of fashion, building experiences to shop for items on Snapchat leans into the augmented reality, image-altering, hyper-visual technology that has become a well-known and much-used hallmark of Snapchat and its owner, self-titled “camera company” Snap.



from Social – TechCrunch https://ift.tt/3vk8uVV
via IFTTT

Engineered earworms on TikTok aren’t that far off from disinfo campaigns

Ever since I read this Bloomberg story about how songs are engineered to go viral on TikTok, I’ve had one thought in my head – if you can call it that – it’s more of a noise, or impression: 

AHHHHHHHHHHHHHHH

Yes, it’s the sound of internally screaming. Just when I thought I understood how deeply social media algorithms have hijacked our desires, tastes and preferences – WHAM! Another jab straight to the nose. I have to admit, I was blindsided by this one. It knocked me out.

Now, I understand that I work for a website called TechCrunch, emphasis on the tech, but if this story doesn’t make you feel at least a teensy bit like a Luddite, well, I don’t know what to tell you. You’re probably like that character in the Matrix, Cypher, who wants to be plugged in.

Is that harsh? I mean, companies are going to company, and partnerships with major record labels is a common sense move for a social media app all about honing the art of short, clever combinations of sound and video. And fair dues to the creators, many of them in college or high school, for jumping at the chance to make some money and get a little bit of fame.

But it’s probably not too harsh when you consider what else is possible when catchiness is weaponized. Here’s what we know: whether it’s internet memes or political slogans or Megan Thee Stallion’s Savage, what drives information dissemination is not the truthfulness of the content or the credibility of the speaker but 1) how easy it is to remember and 2) how quickly it sparks conversations. 

And would you look at that! Those are exactly the variables music producers optimize for today. What the Bloomberg story highlights, inadvertently or not, is how a #1 pop hit and a piece of political disinformation are not all that different, aesthetically. Everyone’s an entertainer.

Now read to the end of the Bloomberg story. Get to the part where it’s revealed that ByteDance (the Chinese company that owns TikTok), in response to threats of a U.S. ban on the app, recruited creators to orchestrate a seemingly grassroots lawsuit against the proposed ban. And I think: damn. Attention really is the most precious commodity in the world. And we’re just…giving it away.

(Cue the internal screams.)



from Social – TechCrunch https://ift.tt/3tRof6g
via IFTTT

Facebook launches a series tests to inform future changes to its News Feed algorithms

Facebook may be reconfiguring its News Feed algorithms. After being grilled by lawmakers about the role that Facebook played in the attack on the U.S. Capitol, the company announced this morning it will be rolling out a series of News Feed ranking tests that will ask users to provide feedback about the posts they’re seeing, which will later be incorporated into Facebook’s News Feed ranking process. Specifically, Facebook will be looking to learn which content people find inspirational, what content they want to see less of (like politics), and what other topics they’re generally interested in, among other things.

This will be done through a series of global tests, one of which will involve a survey directly beneath the post itself which asks, “how much were you inspired by this post?,” with the goal of helping to show more people posts of an inspirational nature closer at the top of the News Feed.

Image Credits: Facebook

Another test will work to the Facebook News Feed experience to reflect what people want to see. Today, Facebook prioritizes showing you content from friends, Groups and Pages you’ve chosen to follow, but it has algorithmically crafted an experience of whose posts to show you and when based on a variety of signals. This includes both implicit and explicit signals — like how much you engage with that person’s content (or Page or Group) on a regular basis, as well as whether you’ve added them as a “Close Friend” or “Favorite” indicating you want to see more of their content than others, for example.

However, just because you’re close to someone in real life, that doesn’t mean that you like what they post to Facebook. This has driven families and friends apart in recent years, as people discovered by way of social media how people they thought they knew really viewed the world. It’s been a painful reckoning for some. Facebook hasn’t managed to fix the problem, either. Today, users still scroll News Feeds that reinforce their views, no matter how problematic. And with the growing tide of misinformation, the News Feed has gone from just placing users into a filter bubble to presenting a full alternate reality for some, often populated by conspiracies theories.

Facebook’s third test doesn’t necessarily tackle this problem head-on, but instead looks to gain feedback about what users want to see, as a whole. Facebook says that it will begin asking people whether they want to see more or fewer posts on certain topics, like Cooking, Sports, or Politics, and more. Based on users’ collective feedback, Facebook will adjust its algorithms to show more content people say they’re interested in, and fewer posts about topics they don’t want to see.

The area of politics, specifically, has been an issue for Facebook. The social network for years has been charged with helping to fan the flames of political discourse, polarizing and radicalizing users through its algorithms, distributing misinformation at scale, and encouraging an ecosystem of divisive clickbait, as publishers sought engagement instead of fairness and balance when reporting the news. There are now entirely biased and subjective outlets posing as news sources who benefit from algorithms like Facebook’s, in fact.

Shortly after the Capitol attack, Facebook announced it would try clamping down on political content in the News Feed for a small percentage of people in the U.S., Canada, Brazil and Indonesia, for period of time during tests.

Now, the company says it will work to better understand what content is being linked negative News Feed experiences, including political content. In this case, Facebook may ask users on posts with a lot of negative reactions what sort of content they want to see less of.

It will also more prominently feature the option to hide posts you find “irrelevant, problematic or irritating.” Although this feature existed before, you’ll now be able to tap an X in the upper-right corner of a post to hide it from the News Feed, if in the test group, and see fewer like in the future, for a more personalized experience.

It’s not clear that allowing users to pick and choose their topics is the best way to solve the larger problems with negative posts, divisive content or misinformation, though this test is less about the latter and more about making the News Feed “feel” more positive.

As the data is collected from the tests, Facebook will incorporate the learnings into its News Feed ranking algorithms. But it’s not clear to what extent it will be adjusting the algorithm on a global basis versus simply customizing the experience for end users on a more individual basis over time.

The company says the tests will run over the next few months.



from Social – TechCrunch https://ift.tt/3dIuZxu
via IFTTT