Posts tagged ‘bot’


It feels like half of Facebook is bot

21.07.2021

Here’s a screenshot from the new members of one of my Facebook groups—actually the only public one I still have. Since Facebook lets spammers join now, we have to block them manually. Their posts don’t make it through to the group as we have safeguards there, too. But I’m not going to let them inflate, falsely, the member count, which in turn will make it harder for posts to reach group members.
   A lot of these bots—they hunt for large groups and their scripts join them—seem to hide under the guise of role-playing for the Pinoy TV series Halik, and they all chat to each other in automated fashion. As Facebook is stupid enough not to recognize the bot activity, you’d think that at least they could see the script at play here, as these accounts are often new, and they set to work joining large groups and pages.
   They don’t recognize them, or, they do recognize them and allow the bot activity to carry on with their blessing. Each one of these blocked accounts was reported, and as usual they were found to be perfectly fine.
   In this screenshot, there were five legitimate accounts. We used to keep the numbers well down because potential members had to answer basic questions, and even some legit people are too lazy to do that. Back then we would see one legit account joining after weeks or months. I think I preferred that, because it kept the spammers and bots away.
   It certainly gives the impression that bots, based on this sample (and others like it since Facebook’s pro-bot policy change arrived with this group), are running at about 50 per cent of the total, which gels with recent research that Instagram is 46 per cent fake (that is, 46 per cent of all accounts are not legit). Nevertheless, I still see far more bots than humans overall: just get yourself into the fake Halik accounts of the Smiths and Montefalcos, and now they’re branching out into other surnames like Montenegro and Buenavista. Thousands, untouched, the only consistent activity on the wasteland that is Facebook.

Tags: , , , , , ,
Posted in business, internet, technology, USA | 1 Comment »


Facebook goes even more pro-bot with change in group policy

02.06.2021

Why are there antitrust or monopoly laws? Why is the usual interpretation of the Chicago School really, really bad for the United States? Umair Haque’s latest post spells it out pretty well, in my opinion.
   Just an idea: let’s not import any of their dangerous ideas into our society, or allow their ever-growing giants to get more of a foothold in our country (and not pay tax here either). Because we have a tendency to kiss their arses sometimes. Just ask Kim Dotcom. Things like their legal precedents are still persuasive here, and with how different their priorities are, we need to place even less weight on them. Let’s not forget the rules we play by here, and that means whomever enters this market has to play by the same.


Speaking of daft decisions on the other side of the Pacific by dishonest parties who have got too big due to what amounts to lawlessness, Facebook has removed the requirement for users to answer questions when they join a public group. These questions were our way of safeguarding the one public group I still look after there, and over 99 per cent of users (no exaggeration; if anything, an underestimate) who attempted to join were bots. I define bots as including any legitimate account running bot software, which I thought was against Facebook’s T&Cs, but not in practice. I still report a lot of them, though unlike 2014 I won’t do them all. I just can’t report thousands that I might see on a single visit.
   I can imagine why Facebook has done this. This way Facebook hides the number of bots from group moderators (as if we hadn’t known of their problems for the good part of a decade), and protects the bots as they continue their activity across the platform. This will encourage even more bots, and as I identified in an earlier post, I see more bots than humans these days on there (and I’m not even a regular user).
   I knew they were liars and shysters so I imagine this is in keeping with that. Cover up just how badly compromised the platform is by bots.
   I haven’t seen much on this change in Facebook group policy, but as changes go, this has to be the most anti-human, pro-bot move they have made in 17 years. No one ever demanded more rights for bots, but here’s Facebook giving it to them.

Tags: , , , , , , , , , , , , ,
Posted in business, culture, internet, New Zealand, politics, technology, USA | 2 Comments »


How to get your Facebook ad account shut down: do something honest

05.05.2021

‘We can’t level, you crazy bastard, we’re in advertising!’—Paul Reiser as Stephen Bachman, in Crazy People (1990)


Signal

You can run ads with misinformation, and you can launch bot nets of thousands of accounts, but what can’t you do on Facebook? Buy ads that expose their tools with which you have bought their ads.
   That’s exactly what happened to Signal when it attempted to run ads on Instagram that illustrated the targeting.
   I don’t believe there’s anything in their T&Cs that disallow this, but Facebook has never been about those. You can breach them as much as you like with running scripts and creating bots, after all. One bastard streamed a massacre on March 15, 2019, which was accessible for some time afterwards; and a year later, eight copies were still on the platform. Facebook “enforces” what it wants to, and that includes disabling accounts that show just how invasive they are.
   I’ve already had a taste of this after I began deleting my ad preferences and exposing how terrible they were. And they probably didn’t like my pointing out that they were collecting those preferences long after I had opted out of their ad targeting (at a time when their own site suggested that opting out meant just that). Now that feature is gone for me.
   It’s what Facebook does. It lies, and even uses those lies to plant software on your computers that never show up in your programs’ list. And, like Google, the timing of when ad accounts are disabled is interesting: it’s not the first time this took place right after you do something that reveals some hard truths about them.
   Personally, I believe Facebook’s preferences are a joke, so Signal may have found their ad account cancelled not because they reminded everyone that the conjurer had a trick, but just how lousy the trick really was. Imagine getting one of these ads and thinking, ‘That ain’t me at all.’ That would get certain Facebook advertisers thinking twice—that is, those who give Facebook’s many bots a pass, and don’t mind that Instagram is 45 per cent bot, 55 per cent human, and don’t mind that their demographic estimates have no basis in reality. I mean, we’re already talking quite a gullible bunch who are doing an activity that’s marginally above setting fire to banknotes in terms of monetary utility, or donating to Jeb Bush’s presidential tilt in 2016.
   Facebook wants to keep as many of them as possible, and they’re taking no chances. You just never know where the tipping point is, when the masses finally decide to jump ship.

Tags: , , , , , , , , , , , , ,
Posted in business, internet, marketing, New Zealand, technology, USA | No Comments »


Facebook whistleblower gets fired; and a workaround for Meizu Music’s inability to find your SD card

19.04.2021

This is a pretty typical story: find fault with Big Tech, try to alert the appropriate people in the firm, get fired.
   Julia Carrie Wong’s excellent article for The Guardian shows a data scientist, Sophie Zhang, find blatant attempts by governments to abuse Facebook’s platform, misleading their own people, in multiple countries. Of course Facebook denies it, but once again it’s backed up by a lot of evidence from Zhang, and we know Facebook lies. Endlessly.
   Facebook claims it has taken down over ‘100 networks of coordinated inauthentic behavior,’ but I repeat again: if a regular Joe like me can find thousands of bots really easily, and report some with Facebook doing next to nothing about them, then 100 networks is an incredibly tiny number in a sea of hundreds of millions of users. Indeed, 100 networks is tiny considering Facebook itself has claimed to have taken down milliards of bots.
   And people like me and Holly Jahangiri, who found a massive number of bots that followed the Russian misinformation techniques, have been identifying these since 2014, if not before.
   Zhang reveals how likes from pages are inflating various posts—forget the bots I’ve been talking about, people have manufactured full pages on the site.
   She uncovered one in Honduras, and then:

The next day, she filed an escalation within Facebook’s task management system to alert the threat intelligence team to a network of fake accounts supporting a political leader in Albania. In August [2019], she discovered and filed escalations for suspicious networks in Azerbaijan, Mexico, Argentina and Italy. Throughout the autumn and winter she added networks in the Philippines, Afghanistan, South Korea, Bolivia, Ecuador, Iraq, Tunisia, Turkey, Taiwan, Paraguay, El Salvador, India, the Dominican Republic, Indonesia, Ukraine, Poland and Mongolia.

   Facebook was inconsistent with what it did, and its own self-interest interfered with it taking action. In other words, Facebook is harmful to democracy, and not just in the US which has received most of the occidental news coverage. On Azerbaijan, Zhang wrote in a memo:

Although we conclusively tied this network to elements of the government in early February, and have compiled extensive evidence of its violating nature, the effective decision was made not to prioritize it, effectively turning a blind eye.

   She was ultimately fired for her trouble, Facebook saying she wasn’t doing the job she had been hired for.
   So if you are going to work for Big Tech, leave your conscience at the door. That blood on your hands, just ignore it. Red’s such a fetching colour when it’s not on a balance sheet.

Little Tech can be troublesome, too. Last year, Meizu updated its Music app after a few years of letting it languish (a familiar theme with this firm), and it was a real lemon. It wouldn’t pick up anything on my SD card, at the location the old Music app itself saved the files. When I could still access the Meizu (English-language) forum, I managed to post a comment about it. Only today did I realize someone had responded, with the same issue.
   I can read enough Chinese to get the phone to do a search for local music files, and the only things it could pick up are what’s on the phone RAM itself, not the card. There’s no way to point to custom locations such as a card (even though there is a custom search, but it applies to the phone only).


Above: Meizu Music will only find music on the phone’s RAM—in this case sound files that come with the dynamic wallpaper and a couple of meeting recordings I made.

   Eventually I restored the old app through the settings, and all was well. It would occasionally forget the album cover art and I’d have to relink it (who says computers remember things?), but, by and large, Music 8.0.10 did what was expected of it.
   Until this last week. The phone insisted on upgrading to 8.2.12, another half-baked version that could never locate any SD card music.
   Sure I could just move the entire directory of 1,229 songs to the phone, but I wondered why I should.
   Restoring the app would work only for a few hours (during which I would try to relink the album cover art, ultimately to no avail). Blocking the new version the app store did nothing; blocking the entire app from updating did nothing. Blocking network access to the Music app did nothing. Essentially, the phone had a mind of its own. If anyone tells you that computing devices follow human instructions, slap them.


Above: I asked the app store to ignore all updates for Meizu Music. The phone will ignore this and do what it wants, downloading the update and installing it without any human intervention.

   I had a couple of options. The first was to make Migu Music the default—and I had used that for a while before I discovered I could restore the Music app. It’s passable, and it does everything it should, though I missed the cover art.
   The other was to find a way to make Music 8.2.12 work.
   There is one way. Play every one of the 1,229 songs one by one to have Music recognize their existence.
   Using ES File Explorer, you head to the SD card, and click on each song. It asks which app you’d like to open it. Choose Music. Repeat 1,228 times.


Above: I finally got there after doing something 1,229 times. As a non-tech person I know of no way to automate this easily. I can think of a few but developing the script is beyond my knowledge.

   Whoever said computing devices would save you time is having you on. They may have once, but there are so many systems where things are far more complicated in 2021 than they were in 2011.
   You may be asking: doesn’t ES let you select multiple files, even folders? Of course it does, but when you then ask it to play them, it ignores the fact you’ve chosen Music and plays them in its own music player.
   And even after you’ve shown Music that there are files in an SD card directory, it won’t pick up its existence.
   It’s at odds with Meizu’s Video app, which, even after many updates, will find files anywhere on your phone.
   For a music player with the same version (8) it’s vastly different, and, indeed, inferior to what has come before.
   How’s the player? Well, it connects to the car, which is where I use it. But so many features which made it appealing before are gone. Editing a song’s information is gone. Half the album cover art is unlinked (including albums legitimately downloaded through the old Meizu music service), and there’s no way of relinking it. European accented characters are mistaken for the old Big 5 Chinese character set.
   The only plus side is that some songs that I had downloaded years ago with their titles in Big 5, as opposed to Unicode, now display correctly. That accounts for a few songs (fewer than 10) of the 1,229.
   I know Meizu will do nowt, as its customer service continues to plummet. I may still file something on its Chinese BBS (the western one is inaccessible and, from what I can tell, no longer maintained by anyone from their staff), but it’s highly unlikely I’ll be brand-loyal. It’s yet another example of a newer program being far, far worse, by any objective measure, than its predecessor, giving credence to the theory that some software developers are clueless, have no idea how their apps work, have no idea how people use their apps, or are downright incompetent. It’s a shame, as Meizu’s other default and system apps are generally good.
   In the future, I’m sure someone else in China will be happy to sell me a non-Google phone when it’s time to replace this one.

Tags: , , , , , , , , , , , , , , , , , , ,
Posted in China, design, internet, media, politics, technology, USA | 3 Comments »


Facebook prefers bots

04.12.2020

If I was on Facebook for personal stuff, I’m certain I could repeat those days where I found over 200 bots per day, but these days I’m only reporting the ones that hit groups or client pages.
   However, I’d say over 90 per cent of the applicants to one of the groups are bots or at least accounts running automated scripts to get into groups to hide their other activity, or to bombard those groups with spam. Facebook has improved its ratio of getting rid of them, but it still leaves roughly half untouched. In other words, if you are running Facebook bots, you’ll have a one in two chance that Facebook’s own people will give you a pass because they can’t tell what bot activity looks like. Little has changed since 2014.


Two Facebook accounts using the same software, it seems, getting caught on a group page. Both were reported, only one was taken down, despite them using the same techniques.

   I thought I’d also grab some screenshots on how automated activity is actually preferred on Facebook, too. I’ve mentioned this here before but here’s an illustrated example from Lucire’s page.
   First up, an automated addition that has come via IFTTT, which picks up the Tweets (also automated) and turns them into Facebook posts. This looks pretty good, and there’s even a preview image taken from the page.

   Let’s say I want to tag the company involved and remove the signature line. Facebook now lets me do this without starting a new line for the tagged business, so that’s an improvement on where we were half a year ago, where it was impossible using the new look.


   So far so good—at least till I hit ‘save’ and the preview image vanishes.

   I usually get the logo only when I feed in a post from scratch directly on to Facebook (assuming Facebook doesn’t corrupt the link and turn it into a 404). In other words, automated, or bot, activity gives you a better result than doing things directly on the site.
   I realize I could add some lines into the code to force the Facebook scraper to seek out the biggest image, but then we’re going into territory beyond that of the average user, and frankly I’m not skilled enough to do it in PHP. And why doesn’t Facebook require that of the bot when it picked up the page to begin with?
   That’s enough for today—I only wanted to illustrate that earlier example as I didn’t do it properly earlier in the year, and give a fresh bot warning. They’re still out there, and I’m betting most pages and groups have inflated numbers where non-humans are messing up their reach—and that’s just fine for Facebook knowing that people will have to pay to get around it.

Tags: , , , , , ,
Posted in internet, technology, USA | No Comments »


Catfished on Facebook? That’s OK, too, they’re there to provide the tools

11.06.2020

I don’t particularly have it in for Google and Facebook. I’m only pointing out the obvious: if you say your policy is x, or your product is y, then don’t deliver us z. Put it into non-electronic terms: if you sell me a car and I put it into first gear, and it instead reverses, then I will complain. And if you look back through 11 years of critique, that is what lies at the foundation of every post about them. Medinge does Brands with a Conscience, Big Tech does Brands without a Conscience. Once they start being honest and levelling with people, then I’ll stop pointing out their hypocrisy.

Speaking of which, a Facebook user calling themselves Barbara Black has taken a photo of former Miss Universe New Zealand Tania Dawson, using Tania’s photo as her profile pic and, of course, catfishing men. You know where this is going: despite numerous reports from Tania’s friends since the D-Day anniversary, including multiple ones from me, nothing has been done. Facebook tells me that there has been no violation of their terms. Some have actually found it impossible to report the fake profile, as their screen fills up with gibberish.

   Yet again it’s Facebook being on the side of the spammers, bots and phonies, as usual, because they have the potential to help their bottom line.
   I can safely say that all my reports of fake or compromised accounts this year have resulted in no take-downs whatsoever, making it far, far worse than what I experienced in 2014 when I said that Facebook faced a bot ‘epidemic’ (I used that very word).
   Very easy prediction for 2020: despite COVID-19, Facebook will have to remove more fake accounts than there are people on the planet. I reckon it has already happened but they won’t admit it. I just don’t know when people will wake up to the fact that this dubious site isn’t serving them, but at least the fakes have got to such a point now that everyday people recognize them: at some point, we will either know someone, or be that someone, who has been catfished or cloned. I’ve been off it for personal stuff for three years and have missed nowt.

Tags: , , , , , , , ,
Posted in culture, internet, New Zealand, technology, USA | No Comments »