No swearing or off-brand comments: AWS touts auto-moderation messaging API AWS has introduced channel flows to its Chime messaging and videoconferencing API, the idea being to enable automatic moderation of profanity or content that "does not fit" the corporate brand. Although Amazon Chime has a relatively small market share in the crowded videoconferencing market, the Chime SDK is convenient for …
Sgt. Pinback : [In a prerecorded diary entry, Pinback is giggling to himself as he speaks to the camera] I went up to Doolittle in the hall today... and I said -
[a subtitle reads "DELETED"]
Sgt. Pinback : - Doolittle... and he said -
[DELETED]
Sgt. Pinback : - and I said, well -
[DELETED, GESTURE DELETED]
Sgt. Pinback : - and he didn't get it!
They can just go Belgium themselves with a Belgium-Belgiuming Belge-Belgy-Bele-Bele-Belgium donkey. You b know it'll never Belgium work & we'll Belgium-well get around them no matter what they Belgium do to stop us. Belgium them with a pointy stick...
This argument through absurdity brought to you by Belgium, Jujunta sunglasses, & the makers of MindBleach.
There is very likely lobbying going on behind the scenes that every platform that allows users to publish comments should be required to use such API.
They need to be quick, because the number of people in power who do not understand technology is shrinking and the young blood will need bags of money to get convinced.
That being said, content upload filters are just around the corner.
When people were not able to speak freely, they started developing their own languages, not understood by oppressors.
We ought to discuss this in a nice friendly pub in Penistone, and then maybe catch the train to Scunthorpe (via Shitlingthorpe), while watching Arsenal play on streaming. Later on, we can visit Titty Ho, Twatt, Wetwang, and Bitchfield.
Note for non UK readers - these are all genuine places in Britain.
No it isn't. This keeps getting repeated by the big companies who get in trouble for not moderating things properly, but it's just a blatant lie. Human moderation is incredibly easy. It's not cheap, because employing people costs money, but that is the only factor preventing them from doing it. What keeps not being mentioned is that you are not entitled to a profitable business. Would paying enough people to properly moderate the likes of Facebook, YouTube and Amazon's tat store mean they are no longer profitable? Maybe. Tough luck. Lots of business ideas turn out not to be viable.
Cuddles,
To be fair, human moderation is hard, at scale. Getting a couple of decent mods together can be incredibly effective, because if they're doing it full time, they can get through an awful lot of work. You can also get consistent results.
Particularly if you've already got a good online community like say El Reg.
Getting a large team of mods together, all of roughly equal quality is much, much more difficult. Firstly you need them to learn the culture of the site you're moderating. With a small team, that's much easier as you can recruit from people who already know it.
Then you have to consider that if you're moderating say Facebook, there isn't even one community. Something like Youtube is easier, but Facebook has lots of different kinds of pages, groups and communities. And covers everything from families organising parties to political parties.
Now you need to get your large team of moderators to be consistent. Now you have to add in multiple languages across multiple countries and cultures. Now you've got a big headache.
I've no sympathy with the big tech companies. They've got the cash to do this, apart from maybe Twitter. But let's not pretend that it's either going to be easy or cheap. Or that they won't make lots of mistakes doing it.
....is this like Facebook's bullshit "AI" filter (algorithm but managers love the term AI)? When a colleague told us at work "I've been suspended from Facebook for a few days. Because I said I had some lovely faggots the other day". The "AI" thought he was using a homosexual slur as clearly America doesn't eat faggots and clearly because American is the only language Facebook's "AI" understands.
From the article
<quote>
you can also use Comprehend for profanity, but you will need to train your own model." Comprehend is used for detecting a social security number.
</quote>
Erm detect a social security number isn’t that just a regular expression on the text input before it is accepted considering it is a standard format
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2023
