##
I strongly suspect that Bluesky “content moderation” and “safety” is following the same trajectory that Facebook did, and that “distribution” excuses it…In case you missed it: a bunch of people on Bluesky posted about the recent assassination/murder of a prominent right-wing American figure with a statement which in Latin* would be: *requi?scat in ur?n?* — and then they had their posts blocked:
https://bsky.app/profile/nathangrayson.bsky.social/post/3lynw7ennws2h
Truthfully: this sounds like the behaviour of a text classifier which — working from a small training set of postings made by challenging individuals — decided to go do a mass-takedown of offending content.
Colloquially: “a bot worked out that some words were ‘bad’ and took down everything containing them.”
Between this and the recent deployments of age verification [in the USA](https://www.theverge.com/news/775901/bluesky-age-verification-us-south-dakota-wyoming ) and [in the UK](https://www.theverge.com/news/704468/bluesky-age-verification-uk-online-safety-act ), I am wondering if Bluesky’s circumstances are so desperate to not run foul of Government attention / regulation / fines, that it’s taken to proactive and deep compliance in the knowledge that “the nerds will be okay, [they can just run up another PDS or implement client haxx and thereby circumvent the controls.](https://gist.github.com/mary-ext/6e27b24a83838202908808ad528b3318?permalink_comment_id=5690428 )“
That’s not a healthy way to approach bad regulation.
*[*] translated to confuse image text classifiers, just in case they’re still being zealous*
#bluesky #censorship #onlineSafety #onlineSafetyAct #surveillance
alecm
npub197…dqg3w
2025-09-15 11:36:24 UTC
Author Public Key
npub197xf7z4rn3f4sy6zm2hf64jftyxmcr8jskwzfrxdzmv6nfn96hnscdqg3wSeen on
wss://relay.momostr.pinkPublished at
2025-09-15 11:36:24 UTCEvent JSON
{
"id": "689cc948d119f53bf986bf630ad4809bffc2a048efce5f8210536f1452cd6d02",
"pubkey": "2f8c9f0aa39c53581342daae9d5649590dbc0cf2859c248ccd16d9a9a665d5e7",
"created_at": 1757936184,
"kind": 1,
"tags": [
[
"t",
"surveillance"
],
[
"t",
"censorship"
],
[
"t",
"bluesky"
],
[
"t",
"onlinesafety"
],
[
"t",
"onlinesafetyact"
],
[
"proxy",
"https://alecmuffett.com/article/115021",
"web"
],
[
"proxy",
"https://alecmuffett.com/?p=115021",
"activitypub"
],
[
"L",
"pink.momostr"
],
[
"l",
"pink.momostr.activitypub:https://alecmuffett.com/?p=115021",
"pink.momostr"
],
[
"-"
]
],
"content": "## \n\nI strongly suspect that Bluesky “content moderation” and “safety” is following the same trajectory that Facebook did, and that “distribution” excuses it…In case you missed it: a bunch of people on Bluesky posted about the recent assassination/murder of a prominent right-wing American figure with a statement which in Latin* would be: *requi?scat in ur?n?* — and then they had their posts blocked:\n\nhttps://bsky.app/profile/nathangrayson.bsky.social/post/3lynw7ennws2h\n\nTruthfully: this sounds like the behaviour of a text classifier which — working from a small training set of postings made by challenging individuals — decided to go do a mass-takedown of offending content.\n\nColloquially: “a bot worked out that some words were ‘bad’ and took down everything containing them.”\n\nBetween this and the recent deployments of age verification [in the USA](https://www.theverge.com/news/775901/bluesky-age-verification-us-south-dakota-wyoming ) and [in the UK](https://www.theverge.com/news/704468/bluesky-age-verification-uk-online-safety-act ), I am wondering if Bluesky’s circumstances are so desperate to not run foul of Government attention / regulation / fines, that it’s taken to proactive and deep compliance in the knowledge that “the nerds will be okay, [they can just run up another PDS or implement client haxx and thereby circumvent the controls.](https://gist.github.com/mary-ext/6e27b24a83838202908808ad528b3318?permalink_comment_id=5690428 )“\n\nThat’s not a healthy way to approach bad regulation.\n\n*[*] translated to confuse image text classifiers, just in case they’re still being zealous*\n\n#bluesky #censorship #onlineSafety #onlineSafetyAct #surveillance",
"sig": "fc7b6219507d36ea921584f56683e7e78723c245bb93485299590bc98de24c6abd90a1b02ac89fb1c8cd70f8146675d51ec5c0e5314c23f7031440a9fc40c44c"
}