05.03.2026 10:58
👍 0
🔁 0
💬 0
📌 0
Thinking about this (again!) re trust and safety infra on atproto and ActivityPub. The reqs for safe, inclusive spaces tend towards similar directions. I wonder if using the Fediverse Auxiliary Service Provider specification (for spam and blocklists) is another example of this kind of convergence.
We're excited to launch our most-requested Bridgy Fed feature: being able to interact with unbridged users!
Starting today, when people interact with you on the other side of the bridge, you'll get a notification over DM along with a "respond" link, where you can like, repost, reply, or block.
And I would add, smaller services that are compliant are still a signal that the DSA is working as intended and remains valuable. If all big social disappeared maybe DSA would be redrawn, but in its current form it still constrains and regulates smaller services.
I think it's important to note Felix pointed out that Mastodon/Fediverse services are not fully exempt, but are exempt from /some/ of the requirements VLOPs have to follow, he also mentioned that most Fedi are in compliance. More details on DSA requirements here: about.iftas.org/wp-content/u...
Who moderates the social web?
In the second post of our 2025 report series, we look at the people behind the work, their roles, experience, and how many are burning out.
Now with three years of data, we’re also looking at trends.
about.iftas.org/2026/01/21/b...
whoops, will delete and resend...
Who moderates the social web?
In the second post of our 2025 report series, we look at the people behind the work, their roles, experience, and how many are burning out.
Now with three years of data, we’re also looking at trends.
about.iftas.org/2026/01/26/b...
This week 54 MEPs signed a letter supporting 'european' social media alternatives. Eurosky/ATproto, Mastodon/ActivityPub are credible options. But the Grok scandal, which prompted the latter, also highlights the stakes - the safety of people, especially minorities (cont. 👇)
For our first launch of the year, we're kicking off with more blocking features!
Rolling out today: ATProto across Bluesky, Blacksky, and any other PDS can apply domain blocklists to Bridgy Fed. Blocks come with if you Bounce, and we even added a handy list of reputable blocklists to choose from.
Any blacksky.app users who get banned from bsky should try logging into the staging.blacksky.community website. Looking to get feedback + identify any lingering bugs. Email support@blacksky.app with those.
Missing data from non-Blacksky hosted accounts + older posts is expected
The 2025 Social Web Trust & Safety Needs Assessment Report is now available.
Volunteer moderators are burning out. Spam is rising. Legal threats are growing.
The 2025 Report reveals where things stand, and what must change.
➡️ about.iftas.org/2026/01/08/t...
The 2025 Social Web Trust & Safety Needs Assessment Report is now available.
Volunteer moderators are burning out. Spam is rising. Legal threats are growing.
The 2025 Report reveals where things stand, and what must change.
➡️ about.iftas.org/2026/01/08/t...
Good news; a judge has blocked the detention of a technology researcher that the State Department had slated to deport.
However, this is just the start of what will certainly be a longer-term challenge. What can we do about this? A thread.
www.nytimes.com/2025/12/25/u...
What does it mean for an online environment to be "toxic" or "polluted" and can we borrow ideas from environmental modeling to make online spaces more livable?
@gedankenstuecke.scholar.social.ap.brid.gy has a post announcing a new CAT Lab project citizensandtech.org/2025/12/onli...
Content Moderation Systems in Languages of the Majority World.
How do content moderation systems work in non-English languages in the Global South? CDT partnered with regional civil society organizations to examine moderation in Maghrebi Arabic, Kiswahili, Tamil, and Quechua across major platforms. cdt.org/content-mode...
"What does it mean to delay the application of the AI Act for the high risk category as proposed? It means deep fakes are still legal. It means generated, targeted medical scams are still legal." — @alexandrageese.bsky.social #EuroskyLive
If you want to post about the @eurosky.social event feel free to use the hashtag #EuroskyLive
That way you’ll show up in this feed: https://bsky.app/profile/did:plc:ooensn4mr5mhznzypvxelfa3/feed/Eurosky-Live
🔴 LIVE https://stream.place/eurosky.social Eurosky Live from Berlin
Apologies for sliding and hijacking this but ROOST is still accepting applications for our Staff software engineer role! We're a v small team so this person ideally has a mix of backend and frontend experience.
Improving the UI of T&S tools is my personal soapbox
roost.tools/staff-softwa...
I so want to just throw divine.video out to the world, but that's not the responsible thing to do. When I realized it's going to get a lot of attention and use, we had to step back and make sure we had systems in place for it to scale both the servers and trust & safety systems.
Today, we're taking our first step into the moderation tools we've been wanting to build for Bridgy Fed for a while now. The reason we're launching this small iteration first is that: (1) it's recently become more urgent for our community, and (2) we want to discuss our ideas publicly. Some...
1/
People with alternative social media icons on their shirts running joyfully away, arms linked, from angry toxic media icons trying to grasp them from a device.
Launching xodus.online ! Building a movement away from toxic social media, towards alternative social media (open social web) - the why, the where-to and the how! Please support our launch ! xodus.online/2025/10/30/l...
At the Paris Peace Forum earlier today, we announced the release of gpt-oss-safeguard in partnership with OpenAI. Now, we’re announcing the launch of the ROOST Model Community (RMC).
roost.tools/blog/a-new-milestone-for-open-source-safety-infrastructure-and-transparency/
"AI in content moderation is far from a recent innovation: machine learning (ML) classifiers have been used for content moderation since around 2016." However, this trend has been accelerating more recently. www.techpolicy.press/researchers-...
Interesting write up.
[New Paper Alert] Who’s Telling the Story of the Russia-Ukraine War on Telegram?
“Shaping Western Perceptions: The Role of English-language Verified Telegram Channels in Framing the Narratives Around the Russia-Ukraine War” by @gruzd.ca, Yisheng Li & @philipmai.com socialmedialab.ca/2025/10/27/n...
Thanks Jaz for joining me on Fireside Fedi. Please check out Jaz and IFTAS . Jaz's work in #trust and #safety on the #fediverse is absolutely vital.
#Peertube #VOD - video.firesidefedi.live/w/rfCpqU15if...
#Castopod #Fedicast - audio.firesidefedi.live/@firesidefed...
www.youtube.com/@btfree_org
Trying out notebooklm created videos trained on my revolution.social podcast. Here's a video about how hard it is to do content moderation, based on my discussions with @yoyoel.com, @noupside.bsky.social, @mmasnick.bsky.social, @coachtony.bsky.social, and others....
