Unable to load image

Bard spotting on Bluesky 09 January 2025 :marseyrandom:

Here we spot wild Bardfinn Bluesky activities.

Be valid and ping ! bardfinn for something worthwhile or create a new thread.

2
Jump in the discussion.

No email address required.

Ms. Penny Oaken, SkyWitch (@skywitches.net):

In one of Reddit's transparency reports, they disclosed that ~94% of 31,000,000 user reports they received in 2021 were false, duplicate, & undecidable reports. 19 out of every 20, garbage.

Dave Willner (@dwillner.bsky.social):

First: Meta will not try to proactively detect "lesser" violations, instead relying on user reports. It is unclear (probably intentionally) exactly what this includes, but it definitely includes hate speech. Unfortunately, user reports are crap, have always been crap, and will always be crap.🧵 2/11


Ms. Penny Oaken, SkyWitch (@skywitches.net):

Meta is moving their reports handling ("moderation") out of California to Texas - a state controlled by GOP theocratic bigots. This means that, for the many millions of US jurisdictions users of Meta products, theocratic bigotry will govern their use of the product.

This is, of course, purposeful


Ms. Penny Oaken, SkyWitch (@skywitches.net):

They can't automate triage / enforcement on these issues because then they'd enforce against theocratic bigots.


Ms. Penny Oaken, SkyWitch (@skywitches.net):

To go back to Reddit —

I tracked hundreds of reports I escalated to second tier after first tier dropped the ball, and found that 25% - 33% of reports in any given category were wrongly dropped as "not violating"; Most people don't know how to escalate to reddit 2nd tier.

Meta, however —


Ms. Penny Oaken, SkyWitch (@skywitches.net):

There is no "Second Tier" at Meta. They have a review board, but it is purposefully a bottleneck only able to handle a very small number of appeals. Once a Meta moderator finds "No Violation" on an item, 99%+ are never reversed. Including obvious violations.


Ms. Penny Oaken, SkyWitch (@skywitches.net):

Meaning that the structural purpose of these changes is to allow hate groups to harass and threaten LGBTQ people and women out of public speech on the platform, without repercussions


Ms. Penny Oaken, SkyWitch (@skywitches.net):

The contractor / employee Moderators will just find "No Violations" on anything that they think falls under the guidelines, so they'll process a report on a dehumanising objectifying misogynist item that opens with the thesis that women are for cooking meals, as not violating - a common phenomenon


Ms. Penny Oaken, SkyWitch (@skywitches.net):

It's called Short Circuiting - humans evaluating speech against a set of rules will pull the lever on the first rule they recognise. They also pull the lever on the first exception to the rules they recognise.


Ms. Penny Oaken, SkyWitch (@skywitches.net):

All of these contractors / employees — especially if they're employed in Texas, an At-Will Employment state — are under pressure to evaluate as many items as possible and are likely managed by their employer firing the least-performing 10% of the workforce every month - a common but legal tactic


Ms. Penny Oaken, SkyWitch (@skywitches.net):

This means these moderators are going to prioritise finding "No Violations" on LGBTQ abuse & misogyny, knowing that they will almost certainly never be dinged for passing these items, but will be fired if they act conscientiously.


Ms. Penny Oaken, SkyWitch (@skywitches.net):

The amount of moral injury to the contractor / employee moderators will be significantly high, as well.

Jump in the discussion.

No email address required.

Charles S (@cstarks.bsky.social):

Is there really anything to this "US-only" stuff? How can they stop people in other countries from seeing content from US users on their network that might violate hate speech rules elsewhere?

Seems like it would be a lot of work to tag each post or comment as acceptable or not for each country


Ms. Penny Oaken, SkyWitch (@skywitches.net):

Facebook has separate moderation operations for the EU & especially for Germany, where open promotion of Nazism is explicitly illegal. But the fact that they're not automating this enforcement & are relying on user reports means that it will leak — it's going to be a deluge of sewage


Ms. Penny Oaken, SkyWitch (@skywitches.net):

They'll rely on user reports from the EU / Germany before limiting visibility of the content there, and that will push the burden of enforcement onto users, who will (naturally) just give up on reporting, eventually, on top of having to field it all in the first place.


Ms. Penny Oaken, SkyWitch (@skywitches.net):

This will, of course, precipitate legal challenges, which may or may not result in any meaningful changes - but will fuel "Europe is censoring free speech" political agendas


Ms. Penny Oaken, SkyWitch (@skywitches.net):

The only reasonable choice is for anyone who doesn't want to put up with extremist hate speech, to quit the platform, or limit their use to private groups and private accounts.

Jump in the discussion.

No email address required.

https://i.rdrama.net/images/16844996389783444.webp

Jump in the discussion.

No email address required.



Now playing: Lost World Theme (DKC2).mp3

Link copied to clipboard
Action successful!
Error, please refresh the page and try again.