There is a new wind sweeping through the online world, Facebook, Twitter and other online social media platforms all seem to be tightening up on abusive behaviour.
I for one am all for this and to some extent support this, I say to some extent because I believe Facebook are using a thermonuclear weapon approach instead of a sniper approach.
A long range targetted single shot aimed a specific target that causes little or no collateral damage and gets the exact job done.
A blanket approach aimed at a very wide area that not only obliterates the target but everything else around it, overkill really.
Facebook appears to have gone down the thermonuclear weapon approach when enforcing their community standards and they fail to understand “localisation”.
I have heard of posts from 5 or more years ago being picked and removed by the “Zuckerbots” (The slang name for Facebooks automatic moderation) which also result in the posters getting some Facebook jail time.
This is happening because Facebook are using an AI to moderate when actually they should be using humans to have the final say, by all means use the AI to search and flag posts which breach the guidelines but don’t allow the AI to sanction people, just flag the breach so a human can make have final say that way they can take localisation and context into account.
In the UK “Fag” is slang for a cigarette In the USA “Fag” is slang for a gay, because the AI isn’t localised it assumes anyone in the UK posting about fags is being disingenuous to gays.
Facebook are modifying the AI as they go but still banning people for days then not apologising when they do tweeak the AI
Anyone that’s ever seen a James Bond type spy movie will understand the phrase “I could tell you but then I would have to kill you” this is a harmless and humorous reference and not an actual threat to kill someone however the Zuckerbots don’t understand this and think it’s a threat to do harm.
Again this needs serious tweaking.
Once a Zuckerbot has taken action there is an appeal system in place but it seems that’s just another Zuckerbot, again this need to be a human looking at it and a human who is localised to the region the post was made from.
Facebook oversight committee
To explain what this is would be beyond the scope of this post but suffice to say it is actually a human derived appeal system but only available after someone has been sanctioned.
The biggest issue with this is it takes about 6 months from the time an application is made until a decision on the case is made, how exactly does this help anyone who has been banned for less than 6 months?
This last step needs to be done within the ban timescale IE if you ban someone for 30 days then you need to hear and decide their appeal within 30 days.
These are just my thoughts on the new wind