June 2017

In December, some of Facebook's policies on hate speech leaked to the public, revealing that in many aspects there approach were, to put it kindly, lax and to put it harshly, bigoted. They've been untangled that particular PR web for more than six months now, but now another leak has set them back even further.

On Wednesday, ProPublica published a new set of leaked documents, this time relating to moderator training. Among all the images revealed, the most damning was a multiple choice slide which gave three options asking which of them were 'subsets [they] protect': black children, women drivers and white men. White men was the correct answer.

The key thing to take away from that is that Facebook define a 'protected class' who are considered immune to extremist or inflammatory speech. In short, you're allowed to say hateful things about certain groups of people on Facebook, but not others.

The documents reveal that groups categorised by gender, sexual orientation, national origin, religion, ethnicity, gender identity or disability are protected. Meanwhile, groups defined by continental origin, social class, age, appearance, occupation or ideology are not protected, neither are religions or nations as a concept (so it's not OK to attack Jewish people but Judaism as a whole is fair game).

That sounds bad enough already but then you find out that a combination of a protected and not protect category in one statement equals not protected. For example, if you were to say "I hate black people", that's an attack on a protected group, so it's prohibited. However, if you were to say "I hate black children", age groups aren't protected, so it cancels it out and suddenly, as far as Facebook are concerned, what you're saying is fine.

This ridiculously backwards logic is, it would seem, the entire bedrock of Facebook's approach to moderation, and once you understand that, their continued failure to properly handle hate speech suddenly starts to make a lot more sense. There are caveats, specifically calls for exclusion, segregation or violence. Once that happens then the whole subset rule flies out the window, but anything less than that still takes this twisted logic onboard.

If you look hard enough, you'll find statements on the platform which could easily be regarded as calls for exclusion, segregation, violence or even all three which easily made it past this flimsy firewall. The perpetrators might not necessarily have even known that they were exploiting a loophole, and in a way that makes it even worse.

Facebook's stance on this kind of thing has long been clear: they don't have any obligation to police it, but these disturbing slides reveal that there's more going on here than just wilful apathy. Regardless of how responsible Facebook may or may not be for the way their two billion users behave, they simply aren't taking it seriously enough.

Of all the overused words in social media jargon, 'discover' is fast becoming the most brutally flogged dead horse of the lot. Snapchat, Twitter, Instagram and Facebook all have their own iterations of it, and now Facebook are adding one to Messenger. I don't know when exactly California was blighted by such a severe thesaurus shortage but somebody should send them an aid shipment before the 1984 comparisons get any worse.

Messenger's version of 'Discover' is at least distinct in its own right - it relates to chatbots, another dead horse which is being pulverised at the moment. While a few bots have been proven as worthwhile, interesting ventures, the lion's share amount to little more than vapid distractions. Despite this, Facebook are resolutely determined to keep pushing them until the outlook improves, and Discover is their latest skirmish.

As the name suggests, it's a tool built into Messenger to help users browse through chatbots and find ones that they might find interesting. Within this section you can search for specific bots, browse categories, and see which bots Facebook are 'featuring'. This addresses one of the biggest issues with chatbots - they have to be sought out.

Before, you'd probably only find out about a chatbot if it was being promoted by a page you followed or a publication posted an article about it, so there was no way of really investigating them to find ones that might suit your interests. That's the gap Discover has been introduced to plug, but more than that, it's there to remind users that chatbots are an integral feature of Messenger, not just a passing gimmick.

Thus far chatbots have largely been, for want of a better word, thick. The promise of interactive conversations with deep learning AI bots was enticing, but there's been little so far to prove that it's any more sophisticated than the chatbots you used to mess around with on MSN when you were twelve.

Now that Facebook have started treating them more like single purpose utilities rather than placeholders for actual people, the outlook is starting to improve. In particular, they're pushing brand interaction, which is why this Discover feature makes so much sense - it removes the need for a Google search when you're looking for an online customer help service. There's certainly more potential beyond that for chatbots, but for now, Discover is a good foundation.

In the increasingly competitive world of dating apps, Hinge has struggled to find an identity. First of all, they boasted that they could use Facebook profile data to offer more accurate matching than any other service. Later on, they threw that out in favour of a format that was better suited to 'serious relationships', which largely involved brutally lambasting Tinder.

Now Hinge are taking another step to stand out from the crowd - they've introduced a video feature. While most of the big social media platforms are all scrambling to provide the most interesting video options, virtually none of the most popular dating apps have it. Back in the days of old school online dating, a video bio was a common feature, but since apps took over it's fallen by the wayside.

As of the latest update, Hinge users can add videos to their profiles, either from their camera roll or imported from Facebook or Instagram. The videos will act as an alternative to photos, so users are still limited to the same six slots and they have to decide for themselves what the ratio between photos and videos ought to be. When users are scrolling through a profile, the videos will autoplay on mute.

The idea with this isn't to create a face-to-camera testimonial, but rather to take pre-existing content and make it part your profile. Studies have shown that people who use photos of themselves actively doing things on dating apps tend to fare better than those who stick resolutely to posed, conditioned selfies. Translating that principle to video adds a whole new dimension to the process, or at least that's what Hinge are hoping.

This may be a significant first step not just for Hinge but for dating app culture in general. Videos are the most popular form of content online, it was only a matter of time before dating apps started to implement them. Hinge's approach may well end up setting the standard for all dating services, a feat which could finally grant them a real identity.

Twitter has a reputation for being the first place to pick up breaking news. The nature of the platform means that when something newsworthy happens, the easiest way to broadcast information about it from the ground is through a tweet, especially in countries with less powerful internet. Now a study has found that beyond that, Twitter is actually a better riot detector than actual police forces.

The study was carried out by Cardiff University, and they used the 2011 London riots as a yardstick. According to the study, it took police an hour and 23 minutes longer to pick up reports in Enfield through traditional means than it would have done if they had been scanning Twitter.

Of course a lot has changed since 2011, and it's now fairly common practice for police to keep an eye on social media. Criminals have been tracked down through Facebook appeals, evidence has been gathered from profiles and, yes, riots have monitored by keeping track of activity on Twitter and Facebook.

The researchers used a set of machine learning algorithms and scanned 1.6 million tweets relating to the riots - which quickly spread beyond London - to create a kind of virtual database, paying particular attention to the timing, location and content of the tweets. Similar tools have been used on Twitter to predict things like election results and even when earthquakes will hit.

The benefits of a 'Twitter police scanner' extend beyond timing, they're also accurate in terms of locating the source of the problem. Currently, the partnership between law enforcement and Twitter is somewhat abstract, each department has their own approach. If Twitter are willing to create a kind of 'one size fits all' system for police and emergency rescue departments could use, it could significantly improve efficiency and even save lives.

Facebook are setting their crosshairs on the younger demographic with the latest update to Messenger. The focus of the update was almost entirely on video chat as they attempt to bring it back into contention with the ever-expanding pool of dedicated video chat apps vying for the attention of the youth market.

Included in the update are filters, augmented reality masks and reactions. Starting from the top, the filters enable users to change the colour palette and lighting of the video feed. The masks, meanwhile, behave in the same way the ones on Snapchat do, overlaying the video feed with various animations which change according to how you move, this can vary from graphics which change the shape of your face to cascading objects like stars and bubbles.

The reactions are tethered to the reaction emojis that you see on the standard Facebook platform; selecting one of them brings up a set of custom filters which correspond to that particular emotion. Basically, they grant you a more visually interesting way than just using a standard emoji. It's a nice addition, but it's hard not to point out that however sophisticated it is, this is video chat we're talking about, and most people are availed of a face. Those are quite good for reacting with.

Finally, there's now dedicated screenshot button so that you don't have to use your phone's in-built screenshot tool if you want to save images of a video chat. Smaller but more practical changes like this tend to be the ones that get the best response, and Messenger has already been criticised by some for being too cluttered.

That being said, it's the teen audience that Facebook are trying to lure in here, and fun features like this are a proven winner among the younger crowd. There are plenty of stripped down, function over form video chat services out there, Facebook have nothing to gain from sticking to that blueprint. The risk of annoying those who already found the app too confusing is certainly present, but Facebook have historically struggled to appeal to teenagers, so a few annoyed old schoolers in favour of a better rep among the fidget spinning demographic probably won't bother them much.

It's been thirteen years now since Facebook first came onto the scene. In that time, it's slowly made a name for itself as the face of social media and racked up a very substantial number of users. 2 billion to be precise, according to a post that Mark Zuckerberg put up on the site yesterday.

Img: TechCrunch
The social network previously hit the 1 billion mark back in 2012, and in taking less than five years to reach its second billion, Facebook has shown that it's not prepared to back down from the competition anytime soon. In fact, new data shows that it has the largest user base for a social platform by at least 500 million, with YouTube and Facebook's own WhatsApp and Messenger the only other apps to have over a billion users.

The rate at which the site is growing is continually increasing despite its age, with the level of traffic it receives also on the rise. Around 66% of monthly users return to use the site day after day, whether that be via computer or mobile phone app, compared to the 55% who used it consistently when the 1 billion mark was reached.

With Facebook now having such an incredible impact around the world, they're at a point where using the network as a force for good is no longer a choice - it's a necessity.

Chris Cox, Facebook's Chief Product Officer emphasised that "there's definitely a deep sense of responsibility in every part of the company" and that they're "getting to the scale where we have to get much better about understanding how the product has been used". With recent controversy surrounding the site over the posting of offensive, violent and illegal content at the forefront of everyone's minds, the company has a duty to ensure that its substantial user base is kept safe and united as they work towards reaching the next billion.

"I live with the constant goal of understanding, for every single thing that we do, how do we maximize all that goodness, and curtail anyway that it can be misused or turned into something sad," were the last words that Cox had on the subject.

Hitting this milestone is another great step forward for Facebook and as improvements in technology and ease of use for the app continue, it's not hard to imagine the site making it to 3 billion users in less than 5 years. While they still have a way to go to ensure that they retain safety on the site as it grows even bigger, now is the moment to take a step back and commend the company for it's incredible achievement.

Well done Facebook!

Al Arabiya
Microsoft, Facebook, YouTube and Twitter have all attended various summits and discussions about counter-terrorism recently. Most of them have resulted in a lot of promises, but few actual results. With pressure mounting against all four companies to do more in the wake of a slew of recent attacks, they've taken another step, but this one looks to be more significant. They've formed The Global Internet Forum to Counter Terrorism.

The forum has been set up to continue expanding various initiatives that the companies have laid out individually, but also to communicate with each other openly on how they're continuing to find new ways to locate extremist content, delete it, ban users and discourage potential recruits from joining up (known as "counter-narrative"). Governments, agencies and global organisations will also be able to use the forum to keep track of developments.

The key idea is to pass ideas back and forth so that all four companies can make sure they're doing at least as well as the other three. Idea sharing isn't exactly a typical habit in the tech world, but when it's in service of something as important as counter-terrorism, exceptions can be made. The forum also plans to work directly with the UN, running workshops to get their in-house counter-terrorism department up to speed on everything they're doing.

The original Internet Forum was launched by the EU two years ago, and the new one is meant to expand on that partnership. Additionally, this will make it easier for governments to actually monitor what the companies are doing to prevent terrorist recruitment, which has been a major sticking point of late.

Tech companies don't like sharing information with government bodies, but at this stage in the game it's impossible to deny that social media has a big hand in the spread of extremist ideas and even terrorist attacks. This new forum could be construed as a kind of compromise, in that sense.


Sharing options on Instagram have always been relatively basic, that is to say, you share a post and all your followers can see it. When the platform first emerged, there was no need for anything more sophisticated than that, it was just an app for sharing photos, but now it's one of the largest social networks out there and it's used for far, far more than uploading a landscape shot of a pretty sunset.

Reports have come out that Instagram are testing out private and/or limited sharing options, so that users can share content only to select groups of followers. This way, users would be able to customise their audience, which could drastically change the way people use the platform. Hackers have been tinkering around with this for years, but nothing has ever stuck.

Instagram have clearly recognised the demand for this feature, and they're taking cues from Facebook to bring it to life. They're calling it 'favourites', and as the name suggests, it's a list of close contacts which can be fully customised at will. When sharing a picture/video, you simply choose between the standard share button and a 'share to favourites' button. There's also a new tab which allows you to view everything you've ever shared to your favourites in one place.

The only way you'll know if you're on anybody else's favourites list is if you see a little green badge above their posts. If you can see that, then you'll be able to visit and view their favourites tab, and if you suddenly find that it's empty, it means that they've removed you, and it's probably time to start figuring out what you need to apologise for. You can't request to be added to a favourites list, either, it's all completely geared towards the user's discretion, which means that most users will likely only populate their favourites list with people that they interact with regularly.

It's more upfront than Facebook's version, and also more basic, as Facebook encourages you to make multiple lists, one of the reasons why their select friends feature has never really taken off. The green badge is also key; much like the little red number you see when you get a notification, seeing a green icon above certain posts will likely provide users with a little shot of dopamine, encouraging them to respond in kind.

Currently, the feature is being very carefully released to a small pool of users, but Instagram have been developing it for the better part of a year, and their enthusiasm about it suggests that they're keen to get it launched as soon as they can.

Conde Nast Travelers
Weirdly, despite less than impressive sales figures, Spectacle's might be Snap's most consistently reliable venture. It's having little to no effect on their credibility or IPO, it's not being ripped off by another other major tech companies and they're not competing with anyone else to promote it. That last one is particularly important, because it means that they don't have to keep up with trends, it's their playground.
That's why they've done things like, say, using giant yellow vending machines to distribute them, and now they've broken out a very interesting new PR stunt - they've sent the Spectacles below the waves. Teaming up with Royal Caribbean, they've developed a kind of specs/diving mask hybrid, effective for 30 minutes and capable of working in depths up to 150 feet.
To demonstrate, three divers took the 'SeaSeekers' for a test drive. The first, marine biologist Gabriela Nava, used them to demonstrate what coral planting looks like. The second, wildlife photographer Roberto Ochoa, used them to search for migrating whale sharks, and although he didn't manage to find any, he caught a lot of other interesting marine fauna in the process.
They saved the best for last though. On Sunday, they released footage of free-diver Ashleigh Baird plunging into Belize's Great Blue Hole, believed to be the largest marine hole of its kind on the planet. It's pretty astonishing, from the descent itself to the stingrays she finds deeper down. More videos are going to be released sequentially on Snapchat and Twitter to demonstrate the capability of the SeaSeekers.
The hope is that eventually the gear will be made commercially available if Snap can get the patent secured. Initially at least, they'll only be accessible to Royal Caribbean cruise-goers, but if that goes well then they might get a wider release. Knowing, they'll probably end up distributing them using a fleet of yellow submarines or something. 

Facebook is the undisputed king of social media news consumption, for better or for worse. The unrivalled user figures, layout and algorithmic approach all lend themselves ideally to news spreading, even if the credibility of said news is sometimes questionable. Twitter still has some weight to throw around and Snapchat is carving out a niche of its own, and now, according to one study, WhatsApp is making a run for the podium.

Reuters international and the University of Oxford surveyed over 70,000 people in 36 different countries, 15% of respondents said that they use the messaging service as a news source. Given that WhatsApp has no home page, news feed or journalistic accounts to follow, the pertinent question isn't so much 'why?', but 'how?'. As it turns out, people like to share news stories in chat groups, to the extent where doing so has become a prevalent source of information. DIY news feeds, if you will.

The percentage might seem small, but the study says that it's now the second most prevalent news source in nine different countries. In particular, 51% of Malaysian respondents said they use it, and while much of South East Asia does exist in a kind of behavioural microcosm where social media is concerned, it's still very telling.

The biggest issue with Facebook (which is actually declining in popularity for news across many nations) is that it has an automated system to decide which news items are worth your time. In that sense, it's hardly surprising that many people are instead turning to groups of people they know and trust to decide what news is worth reading. You could argue that this is an even more confined echo chamber, but all you would need to do to rectify that issue is actually talk about it, flag up the fact that there hasn't been enough challenging content. You get out what you put in.

While Facebook are scrambling to deal with the ever growing fake news issue, WhatsApp is looking more and more like a viable alternative. The same study also revealed that only some 24% of those questioned thought that social media was any good at differentiating between real news and bogus news, so who knows, we might one day be looking at a world where WhatsApp is the most reliable ticker on the market.

At its core, LinkedIn is still a job-hunting tool. They may have added a lot more strings to their bow recently, but if you surveyed the entire half-million strong user-base an asked them why they signed up, the most popular responses would undoubtedly be 'to look for a job' and 'to look for new staff'. Despite all the new bells and whistles, LinkedIn have never lost sight of that prerogative, something which is clearly demonstrated by their latest addition.

Now, when you view your profile, you can see a Search Appearances tracker which keeps track of how many times your profile has been discovered by people looking for new employees. For standard users, you can only see how many times you've been searched, but if you're a member, you can see which specific companies have been eyeing you up, then follow that to find job opportunities and/or contacts within said company.

Even without that, seeing how frequently you're being looked up can be a valuable resource, as it enables you to figure out which profile features stand out most to employers. Give it a few months, or even just a few weeks, and I'll bet any money that we'll be seeing a big influx of advice columns about LinkedIn profile editing based on information gleaned from this feature.

This isn't even the only recent addition which helps people track down work. Open Candidates, which came out late last year, enables people to notify recruiters that they're on the lookout for work, and a more rounded messaging service has been available to all members for several months now.

The nice thing about LinkedIn is that it's a playground for recruiters. The wide range of features means that there's no one approach to job hunting, and depending on what field you're in, there are a lot of different, creative ways to get yourself noticed. This addition not only capitalises on that, it enhances it.

Constant GPS is one of the more controversial side-effects of satellite mapping. It's always optional, but understandably, some people aren't so wild about the idea of other people knowing exactly where they are at any given moment. Still, 'social mapping' is a much sought-after feature at the moment, and Snapchat have doubled down big time to get their own version.

Zenly is one of the more successful, recognisable social mapping apps. It's been downloaded around four million times, and is particularly popular among the teenage crowd. As it turns out, Snap acquired Zenly. Nothing is completely clear, but reports suggest that Snap threw down somewhere between $250 and $350 million for Zenly, and are letting them continue to operate more or less unimpeded.

At the same time, Snap have introduced a feature for Snapchat which is eerily similar to Zenly. Snap Map launched globally for all platforms in yesterday's update. It enables you to share your location with all your friends or a chosen group. Once your location is shared, it stays active for a few hours, unless you go into 'ghost mode'. Even when your location isn't active, you can look around the map to see where your other friends are.

Predictably, Bitmoji also factor into this. When users with a Bitmoji activate the feature, you can see their little avatar standing on the map like a progress marker in an old-school video game. The map also features 'heat' areas, highlighted in red, which indicate areas where a lot of Snaps are being shared. This is meant as an indication that something interesting might be happening, and ties in with the other new feature - Our Story.

Our Story enables users to submit their Stories in the hopes of them being publishes by Snapchat for 24 hours for all to see. In this sense, Snapchat are trying to bridge the gap between social use and current events. Zenly is more designed for helping people figure out when/where to meet up, whether as Snap Maps is about announcing what you're doing and where you're doing it.

Online abuse takes many forms, it can be done purely through messaging or commenting, or it be more elaborate. Some involve misusing profile images, to the point where some people are afraid to share any images featuring their faces online. It's been an issue since the MySpace days, but little has ever been done to combat it up to this point.

Facebook, the most likely place for this kind of abuse to occur at present, are now taking measures to deal with the problem. They've been testing a set of features in India which protect profile images from being misused by other users. As a whole, the set is called 'photo guard' and it stops things like the downloading, screenshotting and tagging of profile images by other users.

The system is optional, and can be toggled on and off directly through the profile image page. When active, the picture itself shows a blue border and a little shield thumbnail at the bottom. This is not only visible to the user in question, but to everyone who visits their profile. It may arguably have been better to keep the shield invisible to others, but in either case it's a positive step.

Facebook are also testing out new ways for people to customise and edit their profile images. Profile filters have been increasing in variety and scope for some time now, and it seems like just about every event, election and international holiday has its own bespoke filter, but apparently more can be done.

They've been experimenting with ways to add different artistic overlays to profile pictures. Such overlays aren't quite as bold or noticeable as the filters which are already out there, they seem more geared towards giving profile pictures more of an artistic flair. Facebook have also said that their research shows images with overlays are 75% less likely to be copied, and also that if someone suspects that their profile image is indeed being misused, the overlay makes this easier to confirm and deal with.

Germany have been very clear on their zero tolerance stance on hate speech. They've leant heavily on Facebook to approach it with a firmer hand, even going so far as to threaten them with hefty fines, and now it seems like they're taking matters into their own hands. On Tuesday, German police raided 36 different homes, all of them occupied by people who stood accused of posting hateful content on social media.

The individuals being raided had mostly been posting right wing extremist content, but allegedly two had been posting left wing content and one had been making threats and prejudiced remarks about the sexual orientation of other individuals. Any more detail than that isn't forthcoming, and you wouldn't expect it to be, but it's fairly easy to fill in the blanks from there.

Laws regarding hate speech on social media differ wildly depending on where you are in the world, but in Germany it can yield a prison sentence of up to five years, if you're found guilty of inciting racial hatred. Even with that in mind though, it's rare that police action, and especially police action this severe, is taken against people who mouth off on social media.

In a sense, this was done to prove that the draft law currently being debated by the German parliament isn't actually necessary. Under that law, platforms would have 24 hours to remove "obviously criminal" material and seven days to decide on anything which doesn't quite reach that margin. Platforms would also be required to provide a fast-acting, easily accessible complaint system. Failure to meet these requirements would result in fines of up to $53 million.

During a hearing on Monday which assessed the draft law, several experts stated that it was unconstitutional, so it may well never be passed. Facebook in particular have acknowledged the fact that they take too long to deal with this kind of thing, but thus far they've only made promises to improve, rather than offering demonstrable results.

For a long time now, Facebook's R&D boffins have been teaching the company's AI bots to communicate with one another. Obviously part of that process involves teaching the bots to speak in human language, but recently something rather peculiar happened. It was found that, left to their own devices, the bots had developed a secret language all of their own, so they were still communicating with each other, but the developers had no idea what they were actually saying.

Before you start digging a bomb shelter, don't worry, they've since figured out how to limit the bots' vernacular so that they can be understood by humans again. Turns out they were still discussing the subjects their human handlers had tasked them to discuss, they'd just been doing it in an entirely new language.

So, what does a secret AI bot language actually look/sound like? Are there a myriad of new words, clauses, tenses and idioms so intricate that they would make Tolkien blush? Well, no, not really. In actual fact, the language consists of a relatively basic set of English words fashioned into phrases like "balls have zero to me to me to me to me to me to" and "you i i i i i everything else . . . . . . . "

Those are two excerpts from a conversation between a pair of bots. It pretty much just goes on like that for several lines, one bot wants to tell the other about the balls, and the other seemingly has no idea what it's prattling on about. Weirdly though, although this particular negotiation didn't go anywhere, other similar ones have actually led to resolutions.

What they're saying to each other can't be classed as a language in the traditional sense, it's too limited, but what's interesting is that the bots are using it to try and resolve tasks in the way they normally would. In the past when bots have developed their own language it's had structure, but lacked meaning. This is almost the opposite.

Getting AI systems to communicate with each other isn't all that difficult, but getting them to learn and develop with only the most minor human intervention is difficult, and it's key to everything Facebook are doing in that realm at the moment. They want their bots to reach a stage where they're not only communicative, but empathetic, where they can understand and react to things almost on human terms. Cyber pig latin aside, they still have a very, very long way to go.

When you think about Twitter bots, the association is more with auto-tweeting than anything else. Bots are a way of signal boosting content to make sure it achieves maximum reach (or in some more unfortunate cases, attacking other accounts by spamming hashtags). Chatbots do exist on Twitter, but they're much more prevalent on Facebook.

There's no reason why chatbots shouldn't pop up just as often on Twitter, but the company have dedicated much more time and effort to other developmental avenues. That's where Sprout Social come in. The company behind one of the most popular management and analytics tools have turned their attention to chatbots, and have put together a product called Bot Builder.

As the name suggests, Bot Builder enables companies to create customer support chatbots for their brands. Said bots will be able to answer relatively basic support questions, verify user identification and help with other small tasks. If they can't resolve the issue on their own, the bots will act as a bridge between the users and actual human customer support providers.

Twitter has always been the best social media platform for customer service, but it's never really had a serviceable automated system for it. Chatbots certainly do exist on the platform, but they're more of a curiosity than anything else. Some companies have toyed with AI Twitter accounts which reply to users through tagging, but as Microsoft unwittingly demonstrated last year, there's still a lot of bugs to work out of that particular approach.

The question is, since customer support chatbots seem like such a logical step for Twitter, why haven't they developed them themselves? They claim that they would rather make Twitter into a 'canvas' for third parties to develop services that benefit them, rather than trying to please everyone and ending up stretching themselves too thin. It makes sense, and Sprout Social certainly have the means and expertise to do this.

The AI Sprout have used for bot builder is very basic at present. It may well become more sophisticated over time, but the main focus is on enabling brands to build the bots into their pre-existing accounts, rather than creating all new ones. Once you get that, only the most essential chatbot functionality is really necessary. There's no word on a full release just yet, but it probably isn't too far off.

Kent Walker (via Rethink Law)
A few days ago, Facebook blogged about their counter-terrorism arm and outlined all the work they're doing to keep extremists from using the platform. Just about the only other people taking as much heat as Facebook for things of that nature right now are YouTube, and wouldn't you know it, they've posted something about it as well.

Kent Walker, Google's general counsel, posted an op-ed in the Financial Times which was later published on Google's blog. In it, he outlined all the measures that Google are now taking to stem the tide of extremist content on YouTube. Given how much advertising revenue the company have lost over the past few months for this exact reason, it's an easy move to understand.

To start with, they've almost doubled their 'trusted flagger' program, increasing it from 63 organisations to 113, and they've also added more 'content classifiers' to their deep learning system. More significantly, Walker has promised that they will be "taking a tougher stance" on content which doesn't directly violate the platform's terms of service, but spread inflammatory, prejudiced or extremist ideas. Said videos will not be taken down, but will appear with a warning.

Perhaps the most important thing outline in the article however is the steps Google are taking towards better counter-radicalisation. They're employing what's known as a 'redirect method' across Europe, refitting targeted advertising in order to find and reach out to potential terrorist recruits and hopefully offer them a different path. While there's still a lot of question around whether or not said recruits will trust what is essentially a form of propaganda, it's a step in the right direction.

The focus on Europe, and the UK in particular is significant. In the past few weeks there have been three terrorist attacks in the UK - one in Manchester and two in London. While social media platforms cannot be directly blamed for any of this, it has been common knowledge for a long time that terrorists use the internet to spread ideas and plan attacks.

Business Insider
Cannes is perhaps the most prominent, significant film festival on the planet. Cannes Lions is something else entirely. Whilst both are certainly star-studded affairs, Cannes Lions is more about elbow rubbing than anything else, as hundreds of digital and tech industry heavyweights gather in the French Riviera to eat, drink and see Chris Martin and Wyclef Jean. It's a meat grinder.

For companies, it's a chance to almost completely eschew the kinds of in-depth technical demonstrations they usually employ and shell out for big, imposing, tangible ad campaigns. Remember when Nintendo launched the 3DS with a small army of attractive women? Imagine dozens of brands all doing that kind of thing at the same time, then times it by 50 and you've got the Cannes Lions experience.

It turns into something of an arms race - who can put on the best dinner, offer the most free booze, book the best act or bring the fanciest yacht - so batting out of left field can be useful if you really want to get noticed. Batting out of left field is kind of Snap's bread and butter, so it's hardly surprising to discover that this year they've rocked up with one of the most distinct offerings - a big yellow ferris wheel.

The company tweeted an image of the brightly coloured behemoth on Saturday, which revealed that there's also a Snap Spectacles vending machine right next to it. The ferris wheel holds up to 84 riders at a time, and provides exclusive Snapcodes which unlock a special Cannes-themed filter. You get two go-arounds, and after disembarking you're given a yellow, branded lollipop. None of these things are especially new (especially ferris wheels), but combined, they present a memorable, different approach to the Cannes Lions 'look at me!' contest.

Snap's primary directive at the moment is standing out from the pack. Endless copycatting by Facebook, unsettling IPO figures and the ever-changing, ever-fickle social media market have put them on the back foot, so they've tried to focus on assuring investors and consumers alike that they exist beyond Snapchat. It might be their central product now, but stunts like this are meant to demonstrate that they intend to move on to bigger and better things.

Snapchat, Twitter, Facebook, Musical.ly!? We've known for a while that there's a second, TV-focused battle happening alongside the ongoing social media arms race, as the big three scramble to snap up broadcasting rights and ink out deals with TV networks for new content. It's been tight thus far, and now there's a new contender, one that likely very few people would have expected - Musical.ly.

The video making/lip syncing network has been going strong for around two years now, with user figures well into the millions, but they're far from a major player. Despite this, they've made deals with Viacom, Hearst and NBCU, and made the promise that their content with stand out from the crowd because it will actually feature an interactive element.

Well, interactive might be strong word, the idea is that users will be responding to these short videos with content of their own, using hashtags to tether the call and the response together. It's quaint, but nothing ground-breaking. What's really worth paying attention to is the fact that Musical.ly haven't put any money down for this content, and in turn it won't be monetised in any way. No ads, no fees.

That may well change in the future, as this is really just a test to see if this kind of original content has the potential to become a central part of the platform, but for the time being it's certainly a respite from the ad interrupt laden worlds of Facebook and YouTube.

The shows are already being rolled out, with four already available. MTV have released two - shorter versions of Nick Cannon Presents Wild N' Out and Greatest Party Story Ever, while fashion mag Seventeen have produced the other two - 'Fashion to DIY For' and 'Seventeen in the City'. NBCU are working on one which will be released under their E! Entertainment banner.

All the shows are planned to run at 2-4 minutes per episode. Each new episode will appear in the trending section for 24 hours, but even after that they'll be accessible from the profile pages of the companies in question. Users will be able to respond, and view other responses through the app's hashtag page.

Musical.ly has a very young user base (too young, some have argued), and through that they've been able to carve out their own niche. The scale of their growth has been somewhat underestimated, and the fact that they're encroaching on territory which had formerly been dominated by the big hitters suggests that they've got big plans lined up. While Musical.ly might never experience the boom that Snapchat did, it's certainly a platform to keep an eye on.

Social Barrel
One of the things Facebook takes a lot of heat for is their perceived inability to track down terrorists on the platform. They've been pressured by numerous national and international authorities to do more, and almost every time there's a terrorist attack, the suggestion rises that they could have done more to prevent it. Generally, Facebook's response has been to point to new AI developments and simply assure people that they're working on it, but now they've revealed more details.

In a blog post last week, Facebook reveal information about how they find terrorist accounts, and in particular about the actual, human staffed counter-terrorism team they have on staff. Said team apparently contains more than 150 people, many of whom come from law enforcement backgrounds, rather than tech. Facebook claim that 30 different languages are spoken within the team, and beyond law enforcement, it also includes former prosecutors and an emergency response team.

Facebook have also pointed towards their partnerships with national governments, law enforcement authorities and, in particular, NGOs. Facebook have worked with NGOs to develop counterspeech, a kind of anti-propaganda that presents a more preventative approach to counter-terrorism - stopping the ideas being spread before any threat of violence has a chance to emerge. Their co-operative summit with Google, Microsoft and Twitter is also mentioned.

Of course, it wouldn't be a Facebook open letter without at least some mention of AI, and sure enough there's a section dedicated to it. In this case, it details the workings of their image matching and language recognition systems, as well as 'clusters' - content which isn't directly connected by any one group or page, but all relates to the same extremist ideology. Facebook also claim that they can detect and remove accounts made by repeat offenders much more quickly than before.

This blog post is a fairly direct response to increasing governmental pressure to take further action. Several national governments have laid out plans to fine Facebook for failing to act against terrorists and others have vowed to place stricter regulations not only on Facebook, but on the internet as a whole. For many, Facebook's counter-terrorist wing won't be enough to satiate them.

It doesn't seem like that long ago that Twitter was last remodelled, but wouldn't you know it, they're at it again. All versions of the service - mobile, web and desktop - have been changed, and while no new features have been added in, they've done a lot. Many of the fonts, colours and shades have been changed, settings functions have been shuffled around and all the icons have changed shape. In short, you're going to notice.

If you're an iOS user, however, there's one thing in particular which will stand out. Previously, there was a profile icon on the bottom nav bar; that's been removed, instead you swipe right on the home screen to access a page with your profile, account switching and a small cluster of privacy settings. This also means that the 'profile settings' gear icon has been removed, as there's no longer any need for it. This is certainly neater, but it'll take some getting used to.

If any of that sounds eerily familiar, it's probably because you're an Android user. This version of the app has been the standard on Android for almost a year, and its proven success rate is part of the reason it's being migrated to all the other versions of the platform. All the other changes are completely new across the board, however.

Aesthetically, the most noticeable change is the alteration of profile icons, which are now circular instead of square. It might not seem like much, but it's been in the pipeline for a while, and seems to have been inspired at least in some capacity by Messenger and Instagram. The reply icon has also been changed, the arrow has been ousted in favour of a speech bubble, something which gives you a clearer insight into Twitter's current ethos.

If you're OG, you know that the arrow denotes a response, but anyone new to Twitter might be confused by it. A chat bubble, meanwhile, is far more universal, suggesting that Twitter are trying to appeal to newer users first and foremost. One of Twitter's biggest issues is the tendency for new users to fall out of it almost immediately. They'll post a few times, but for whatever reason (diminishing returns, confusing interface, difficult finding accounts to follow), they'll just stop.

Compared to Facebook and Instagram, Twitter's daily active user figures are dangerously low, but in trying to remedy this, Twitter are putting their unique appeal at risk. There's a strange kind of dedication to the service among long-term users, to the point where even changing one icon is risky. This is probably why they've left all the other icons well alone, except to give them a bit of a facelift.

Later is one of the most useful digital marketing tools for Instagram. It started in 2014, and allowed brands and influencers to schedule posts across multiple accounts in much the same way as Hootsuite or Sendible do for other platforms. Since then, they've added in aggregation and audience building tools, gradually morphing Later from a useful service to an almost essential one.

Now they've added in a new feature which may well be their most useful yet  - one which shows you when the best time to post is. Since Instagram changed the way posts are ordered in the main feed, engagement has been steadily declining, but weirdly timing is more important than ever. Despite posts no longer appearing chronologically, there are certain peak posting times in different time zones during which more users, and particularly those likely to engage, are active.

What Later can now do is show you these time slots and recommend the best time to post something. It's a tailored system, Later will actually scan your Instagram account to determine your audience, and when the best time to post to them is. Those times will appear highlighted in the calendar view. That's pretty much it.

It's simple, but it homes in on a key issue with Instagram, because each audience is personal, figuring out when it's best to reach them turns into a perpetual game of trial and error, and if you're lucky enough to get some growth, it only becomes more difficult. Even with the feature active, you won't be able to see the best slots for each day, you might get three on one day and none on another. Also, if you're posting at around the same times every day, the feature won't work, because it won't be able to differentiate the data.

Additionally, it will only work if you've posted 50 times in the past six months, but if you haven't been posting that frequently, it hardly seems worth it to use Later at all, especially given the price tag. The feature is only available to those using the $16 per month version of the service or higher, so if you're not using Instagram for business in some sense or another, it's not worth it. That being said, it's certainly cemented Later's position as one of the best Instagram tools on the market, if not the best.

Instagram are trying to widen the gap between normal content and paid-for promotional content - the kind that gets put out by brands and celebrities. Previously, the only thing that really differentiated the two was the nature of the content (which admittedly rarely makes it hard to tell them apart), but now paid-for posts will feature a 'Paid partnership with...' label.

While it might seem like a logical addition, Instagram aren't necessarily doing it because they want to. In actual fact, they've been in hot water with the Federal Trade Commission. A recent report revealed that 93% of the posts made by Instagram's 50 most popular users didn't adhere to FTC guidelines. The main stipulation is that all paid-for content must clearly indicate that it's sponsored, but many influencers sidestep this, as it can be detrimental to a post's popularity. The less corporate content seems, the better it stands to do.

In particular, the FTC warned users that the #sp hashtag was nowhere near clear enough an indication of sponsored content, but that's kind of the point. Users don't search for hashtags which reference sponsorships, but terms like #sp and #partner have a kind of 'wolf in sheep's clothing' effect. Instagram have clearly buckled to pressure to rectify this, which is why they've brought this new label in.

The label appears above the content, which means it's one of the first things users will see when viewing the post. The issue is that Instagram cannot directly enforce the user of the label - it's still very much an 'opt-in' feature, so at present there's nothing to stop users from simply ignoring it and carrying on as they have always done. That being said, Instagram are also opening up metrics on sponsored posts to users and brands, so it will be clear to everyone if sponsorship labelling is really detrimental to a post's performance.

The creators of the posts will be able to view metrics from inside the Instagram app, which creates an incentive to use the label which will, in theory, eclipse the desire to sidestep it. In theory, there's nothing wrong with brands stating their intentions, but advertising and dishonestly have been bedfellows since the dawn of time, to the point where misleading consumers is just force of habit. By offering brands the chance to track their metrics more easily, Instagram have introduced a canny trade-off for better transparency.

Facebook's most polarising feature is being altered yet again. It's been working overtime recently in the wake of several incidents in the UK and Sweden, most recently the fire Shepherd's Bush. The most noteworthy change they're making is the ability to raise funds directly through the feature. In this way, people can take action directly and immediately.

This is significant because before, Safety Check was only really useful for people who wanted to assure others of their wellbeing, but given that you're far more likely to be hearing about a crisis than actually in it, there's little you can really do with that information. This way, you can actually take direct action. Personal fundraisers have been a component of Facebook since March, but this is the first instance of them being integrated with another service.

Additionally, if you do mark yourself safe, you can add a bit more detail about the situation, so it almost acts as a more situationally specific status update. People outside of the situation can also access more detailed information about the crisis, eliminating (or at least lessening) the need to check external news sources when you see the first influx of safety checks.

As always, the feature is being released on a rolling basis. Currently it's only available in very select areas of the US. Of course, we don't know when the next major disaster/crisis will be, so it's impossible to know when we'll be able to gauge the effectiveness of this update.

Fundraising is a big profit area for Facebook, cynical as that might sound. They're currently underselling GoFundMe by a significant margin, and the more convenient they make it to set up and donate within the platform, the more they'll eclipse them. Crises always have a kind of 'call to action' knock in effect which causes fundraising campaigns to spike, something which Facebook are using Safety Check to capitalise on.

On a less callous note, Facebook are trying to make Safety Check more comforting, as past reports have suggested that it's done more to cause fear than to assure people of their loved ones' safety. Another recent criticism has been that it's more of an attention grab than any kind of practical safety measure, but once again the ability to reveal more detail combats that. In addition to all this, Facebook have also made their Community Help service desktop compatible.

Campaign for Liberty
Donald Trump is a very, very active Twitter user, and he doesn't mince words. Since taking office, his often controversial Twitter activity hasn't receded an inch. If anything, it's actually increased, and his old habit of quietly erasing past tweets when the thinking behind them falls into sharper question hasn't gone anywhere either. Well, a new bill is under evaluation which may make it illegal for him or his team to do so.

That's fascinating in and of itself, but then you find out what the bill is called - The Communications Over Various Feeds Electronically for Engagement Act. Shorten that, and you've got The COVFEFE Act. I am not even remotely joking, somebody actually sat down and figured out a way of turning Trump's weirdest Twitter gaffe into an acronym about making it illegal for him to delete his Twitter gaffes. What a time to be alive.

The bill, put into motion by transparency caucus co-chair Michael Quigley, would make all presidential social media posts a part of the presidential record, falling under the 'documentary material' banner. That would include not only tweets, but DMs, Facebook posts, Instagram posts and everything in between. Said posts could still be deleted, but it would be completely redundant, as they would have to be put into the public archive anyway.

This is particularly important for Twitter, because the Trump administration actually use it to clarify stances and policies. It just happens that the man himself also uses it to do things like lambasting the Mayor of London for allegedly not responding strongly enough to the recent attacks in London Bridge and Borough Market. These are the things that his cabinet would probably rather didn't make it into the record, but Quigley is proposing that they do.

Trump has deleted 18 tweets since taking office, some of them for mistakes, and some which made big, sweeping politically charged statements, such as a call for a former assistant US attorney to be subpoenaed. The White House currently has an informal agreement with the National Archives to preserve deleted social media content, but seemingly the National Archives would be left with little to no recourse if The White House violated this agreement.

If this does become law, it will likely result in much more careful monitoring of the presidential Twitter page, but Trump has always done more or less whatever he wants with it, and stern warnings about things ending up as public record probably won't stop him. At least we can take some solace in the fact that it won't be hard to demonstrate to future generations what kind of president he was.

Facebook Groups are certainly useful, but the feature has always been somewhat spartan. You can decide who joins, but then it's just the same as the normal news feed, only in a closed circuit. It could almost be argued that you could achieve the same effect by simply limiting the reach of your regular posts. There's certainly always been room for improvement, and it seems like Facebook were aware of that.

They've been quietly prepping a new version of Groups on the mobile app, and now it's ready for testing. Groups will now be visible alongside the News Feed and Notification sections, and the page itself has been updated with a 'Discover' section which points you to new groups that you might be interested in, based on category.

Alongside 'Discover' is 'Create', which does what it says on the tin. The actual Group creation process is largely the same, but the design of the interface has been given a major polish. Previously, you had to really dig into the settings on the app to find Groups, so even the act of placing them front and centre like this is a significant improvement. Facebook actually tried to branch Groups off into a separate app in 2014, but it fizzled out in a matter of months.

Like almost every recent change that's come to the platform, this comes on the heels of Mark Zuckerberg's gargantuan 6,000 word open letter. Specifically, Zuckerberg discussed "meaningful social infrastructure". The emphasis has certainly been angled more towards letting users decide what interests them, rather than leaving Facebook's algorithms to figure it out for themselves.

You don't generally add people on Facebook because you have similar interests, but the News Feed and the friends list are becoming increasingly separate entities, and there's no reason why it shouldn't be possible to retreat to havens of mutual interest outside of the usual Facebook experience. At present, Groups are more commonly used to organise parties or arrange the office retreat, but this way they might take on a more Reddit-esque format.

Spectacles remain one of the best tactical moves Snap has made - it was the launchpad which helped them move beyond their digital beginnings and towards redefining themselves as a company. While Snapchat has left them facing the looming spectre of financial turmoil, Spectacles have remained successful, especially now that they've gone on sale in Europe. And a second type appears to be in the pipeline.

There's been no official statement yet, and it's a closely guarded secret, but reports from within the company suggest that the new specs will be 'quite different' from the originals, and the fact that Snap filed a patent for augmented reality glasses back in 2015 supports the notion that the new specs will have some sort of AR functionality. Snapchat already has an AR function in the form of World Lenses, so applying the same logic to Spectacles would be an understandable next step.

Snap have been linked to a few other AR hardware blueprints, such as a helmet and a visor, but this one seems the most viable. The current specs can only really record video and take pictures, which isn't all that impressive in and of itself, but now that the design is out there, there are plenty of ways Snap could improve the formula. The new specs might even have updatable software which allows for new 'apps' to be installed, but that's just speculation.

Spectacles have yet to heavily contribute to Snap's earnings - they've like only just broken into six figure sales numbers for them, but as an opening salvo into the world of hardware sales, it's significant. The real aim was to demonstrate that Snap could make, advertise and sell hardware which had the same kind of flavour as Snapchat, and on that count they succeeded. The first Spectacles were almost a gimmick, but the follow-up model may well carry more substance.

In any case, it's unlikely that we'll be seeing any definitive information about the new Spectacles for a while yet, much less a full public announcement. Snap will likely wait until the original pairs have had a bit more time to branch out into more global circulation, and they may even announce some other kind of hardware first (they're allegedly also working on a drone), but any doubts about the company's viability as a hardware manufacturer have been more or less extinguished. 

It's hard to derive much entertainment from social media at the moment. Facebook and Twitter have increasingly become warped sounding boards for the state of the world and recently things have been pretty bleak. 'Active' social media use has faded away and been replaced with endless, virtually pointless scrolling as you rifle through your news feed, barely engaging with the material on offer, most of which is advertising.

The solution is obvious - stop scrolling, either quit social media or find a new way of using it actively, but there might be a third option. What if you could take the actions - the scrolling, liking and commenting - and isolate them from the outside world, turn the whole experience into a kind of release? Well, that's what Binky does.

It started out as a kind of satire, a social network which offers the same feed of content, and the same ways to engage with it, but restricts it solely to you. In practise, it looks a bit like a cross pollination of Tumblr and Facebook, but all the 'Binks' are just things that exist. Let me give you an example, currently the top items on my feed are Jeff Beck, The X-Files, Taoism, the sheepshank knot and a ferret. I can either like, comment on or 're-Bink' these things, but it does nothing.

Well, not nothing, but not a lot. If you 'like' a post there's a small explosion of little stars, and if you re-Bink it there's a cascade of thumbs up symbols. If you comment, it doesn't matter what you try to type, Binky will start just putting words in for you, usually something along the lines of 'omg i love this so much'. It doesn't matter though, nobody else will ever see it. Oh, and you can also swipe left or right, prompting either a kind of fireworks symbol or a big middle finger to appear.

It's amusing, but some have observed that it could actually be the ideal social network for the current climate - it satiates the desire to scroll, like and comment without affixing it to anything in reality. It takes the most addictive aspect of social media and puts it in a closed circuit. If you're doing it on Binky, you might not feel so tempted to check your Facebook every five minutes.

If you really think about it, using social media to procrastinate or alleviate boredom doesn't make a whole lot of sense. It's a tool for communication, for connecting with people, but the instant gratification it provides has clouded its purpose. Perhaps if we all turned to Binky when we were bored, we'd start to remember the appeal of Facebook and Twitter, and start enjoying them again.

Venezuela is in a state of severe turmoil. Citizens are protesting despite heavy governmental suppression and many of the demonstrations have turned violent. This is all a result of movements by the ruling governmental party to dismantle the opposing one, through arrests, silencing and ultimately the dissolution of the party. Millions of Venezuelans have taken to the street to make their anger clear and more still have taken to social media to amplify that anger.

Increased instances of Twitter hacking during times like this are far from uncommon, but now Venezuelan accounts are falling victim to a new type of hacking, which has been dubbed the 'DoubleSwitch'. It involves hacking into a given Twitter account, changing the handle and then opening a whole new account under the old one. In doing this, the hacker will send the owner of the account to the wrong place when they try to recover it, and make it almost impossible to track down their original one.

One notable Venezuelan activist and another journalist have already been hit, and in both cases the hackers used their accounts to spread misinformation about the nation's political climate, and the protests themselves. They would use the new, imposter account to do this, despite the fact that previous followers don't carry over. Even with that in mind, the name and the nature of the tweets still creates a lot of confusion.

Thus far the technique has only been observed on Twitter but security experts have said that there's no reason it might not start happening on other platforms, particularly Facebook and Instagram. Facebook implement a different, arguably better security system, but it's far from airtight and creating 'imposter' accounts is entirely possible. Coping with name changes is new ground for both companies, and there's nothing to stop a name being registered under a different account name once the original has been compromised, it's a pretty massive loophole.

For activists, especially in places like Venezuela, Twitter is an invaluable resource. Hacking instances like this could majorly disrupt their lines of communication. Twitter never asked to be responsible for anything like this, but they've yet to even address the fact that it's happening and that needs to change.

"So here's how I work, I'm going to ask you about your mood and as I get to know you, I'll teach you some good stuff." - Not a phrase you'd expect to hear from a machine. Therapy relies on empathy, and last time I checked AI wasn't capable of that yet (frankly I'm not sure I want to live in a world where it is), so the idea of a therapy chatbot might seem pretty absurd. Well, it's happened anyway.

The above quote is the first thing 'Woebot' says to you upon connecting through Messenger, and then it starts asking questions. Through this, it helps you to 'recognise patterns' and leads you through what is essentially a CBT (cognitive behavioural therapy) course, albeit a lot more basic. According to the researchers at Stanford who tested Woebot, it only takes two weeks of daily five minute check-ins before measurable results start coming in.

Impressive as that is, a robot is no substitute for a real living, breathing therapist, something which Woebot clarifies before you start the sessions. It's meant more as a companion to actual CBT than a complete course. Speaking as someone who has undergone CBT for real, it does play out like a kind of watered down version. You get asked the same questions about you mood and thoughts, and the bot sends videos and tips about daily attitude designed to help shoo all the negative thoughts away.

I did one session, during which the bot sent me a video about the significance of language when describing emotional states. It wasn't the most well-produced video in the world but it got the message across. The more significant thing was that the bot asked me about me mood and offered to extend the session if I felt the need to talk things through further. Given that most Messenger users are active daily, and use it on the move, this is a smart approach.

Once again, this is not a worthy trade off for a real therapist, but even being able to vent your issues with a chatbot has its merits. They haven't been around for long enough for any real, peer-reviewed evidence on their effectiveness to come out, but it certainly can't do any harm. If you're feeling anxious or depressed but aren't sure about taking it further, Woebot might be an effective first step.

Author Name

Free Gift

Free Gift
Get immediate access to our in depth video training on the click by click steps required to get your successful online business started today

Contact Form


Email *

Message *

Powered by Blogger.