Baa baa bad sheep? He’s the naughty sheep of the family?
In IT, permitting or denying a program is now allow-listing and block-listing. Language changes as society changes - but removing the stigma associated with ‘black’ will require major changes in both language and society.
There is no way that an algorithm would rely on simple lists of proscribed words. No way, none.
It would look for contextual meaning.
For the case of advertising wool.
‘Various colours, white, black, green, …’ would be pefectly fine.
‘At the moment, we can send white wool only’ should be perfectly fine as well.
‘Normally we send various wool colours, but at the moment it is whites only’ may trigger some flag for racial reasons.
Meta are a high-tech company, and would have plenty of serious Computer Science people developing the algorithms. Allow lists, block lists? They would laugh hysterically if those types had a sense of humour.
I fail to understand how “Poms” could trigger anything in any context. Even if taken literally what I said was completely inoffensive in or out of context. It HAd to have been the word.
Sometimes posts can be seen by more people than you realise. Somebody could have flagged your post as offensive.
I lived and worked in London and not many there liked being called a Pom. I didn’t really like being called a Convict, but shrugged it off. I once mistakenly called a Cockney a Scouser and just barely avoided a hospital visit.
Depends on your settings and your friends list, and also your friends’ settings and their friends’ lists as to who may see your post and when. A friend ‘liking’ your post could see it go to all their friends.
You really do need to be carefull and think, could someone, somewhere, not like an innocent Aussie term.
remind me not to say “thongs” then!
It’s been done. Apologies if I get the participants wrong, but I think it was Andrew Denton who asked Wendie Malick from Just Shoot Me if she knew what thongs were - and her answer was “I’m wearing one now”.
Good advice. Even calling someone an Aussie might cause serious offence, especially if they are a Kiwi.
But overall most Kiwi’s are very relaxed about most things.
Does ‘FacePlant’ or any other social engineering enterprise have sufficiently sensitive AI to reliably identify the nature of the relationship between two parties online exchanges?
That gives a false impression though. These days it is not the case that “language changes”. People are being forced to change what they say or face cancellation. The changes are just being made forcefully.
It is true that language can change as an emergent behaviour (e.g. some of the crap grammar that passes for correct today) but this isn’t it.
Yeah, nah.
There are even sillier examples in IT but we are drifting too far off topic.
It’s also changing history?
Obliteration of past usage of language seems to be rewriting how our parents, grand parents, great grand parents …. thought and spoke.
Do all examples of language need to be considered in it’s full context, before censorship or a warning applies?
It would seem that Facebook and others AI is not that capable. For many examples the context depends on the period in time the word/s or phrase etc are connected with.
Some examples of problematic language today such as a ‘gay old time’ from the Flintstones are very contemporary.
Other examples of writings from the earlier 1900’s and recognisable prominent Britains/Australians make more than derogatory reference to some of the population purely on race. The content is typically grossly offensive, however most AI is unlikely to detect any concern. It depends on whether any of the writings use commonly accepted racially negative words. Sentiment can be conveyed by more than a single term and escape detection. (For those wondering or looking for examples, SBS series Every Family has a Secret, now up to 3 seasons/series often stumbles across treasured writings that would challenge modern sensibility.)
The next question to consider is whether Facebook et al will next begin to censor the works of Shakespeare or Chaucer and remove from history references of Christians being fed to the lions by the Romans?
The simple response is we cannot change history, nor should we permit a modern pariah such as Facebook use it’s algorithms to hide truth, modern or not so recent. It’s how language is used, not only the words chosen that should be tested?
John Buchan is a prime example. Read The Thirty Nine Steps or any one of his other novels and you will find racial views that are very definitely not in line with modern sensibilities. I did not even notice them when I first read his books - I was too young to understand such things - but revisiting them makes clear the author’s views on race.
Actually, that’s largely propaganda. Christians were not chosen as specific targets to provide feline haute cuisine, and while they were persecuted by Roman authorities this was done in a rather haphazard fashion. References to it should not be removed from history but should only get the level of attention that is appropriate - remembering that almost all of our written knowledge of the Roman Empire has at some point been through a Christian establishment whose scribes would not have been entirely disinterested in Christian martyrs.
Tell that to Richard Wagner, or Adolf Hitler, or Josef Stalin, or Vladimir Putin. History is written by the winners, just as war crimes are defined by the victors.
I have attracted Facebook’s algorithmic censorship for a reason I haven’t seen in the lengthy exchanges above: posting “spam”.
What I posted was my own photograph of an electricity substation in a nearby nature reserve, made up in the image of the famous TARDIS telephone box from Doctor Who, and I posted it as part of a comment under a similar photo of a similar electricity substation a few hundred kilometres away, a photo taken by an old friend who used to live near me and put up on his Facebook page.
He attracted no sanction whatsoever from the Facebook censors.
Facebook threatened all sorts of dire consequences should I persist with my evil “spamming”, and while it offered me the opportunity to dispute its “judge, jury, and executioner” conclusion, endless attempts on my part to do so were met with no response.
Hmm. So what did you actually say in your comment? Think about that. Forget about the image unless it was somehow offensive. You say it was yours, so you own it.
How many ‘friends’ do you have who will see your posts, and how many ‘friends’ does your friend have who will see your comment post on their post?
It can get exponential, and all it takes is one person in a grumpy mood that you do not even know to hit the ‘report’ option and it comes back to you.
As a general rule I do not post images in replies to posts. If someone wants to see some picture I have I messenger it to them personally.
.
I compose my comments in a text editor which saves automatically, so I know exactly what I said.
"Doctor Who’s TARDIS must’ve re-materialised there after dematerialising on Red Hill – I took this picture some years ago, a few hundred metres towards home from the lookout up the top.
"(I believe it houses some electrical network equipment!)
"And as for the gender of the good Doctor, rumour has it that’s about to change (again):
“https://www.bbc.com/news/entertainment-arts-57940451”
Perhaps I offended a Doctor Who fan, but why would they “report” me as sending out spam?
I wondered if the Facebook algorithm saw two photos of the TARDIS, with the second one accompanied by a link to a BBC story, and somehow thought it was part of a publicity stunt from Blighty.
I have only a handful of Facebook friends, ones I can trust to see highly-defamatory stuff without ratting me out, but my friend may well have hundreds – all Facebook will tell me is that we have four friends in common.
Seems all very innocent.
I, like you, keep my direct friends list short. However one of my friends has many hundreds of friends, and has a habit of ‘liking’ lots of posts.
So all those posts were flooding my news feed. Ended up blocking lots of posts from posters I had no relationship with.
I felt like unfriending him, but didn’t. I felt like reporting as spam some of the posts, but didn’t.
Aside from being “done” for other offences I have also been done for spam. I was trying to reach out to quite a number (50 or so) members of a private community group I moderate, by messaging rather than a notice on the page which some will overlook. Because the wording of each message was the same, it was blocked as spam, despite the messages all being from moderator to group member of a page the public cannot access. FYI - I am not really expecting any solutions!
Unless those 50+ members were all ‘friends’, Facebook automatically considers those messages to be spam.
Now if you pay Facebook, it seems you can bombard non-friends with posts without too many problems. Sponsered content.
Not sure if the same applies to messages.
