Daniel Kelley
banner
danieljkelley.bsky.social
Daniel Kelley
@danieljkelley.bsky.social
Gaming and tech advocate; Brooklyn native; Director, Strategy and Operations, Center for Tech & Society, ADL
I spoke to @willoremus.com of the Washington post on how giving a megaphone to Nick Fuentes is a terrible idea and is a case study on how backsliding on content moderation by social media companies is opening a door to bad actors like him: www.washingtonpost.com/technology/2...
Far-right provocateur Nick Fuentes is triggering a MAGA civil war
Once barred from nearly every social media platform, Fuentes’s resurgence has driven a wedge through the conservative movement.
www.washingtonpost.com
November 9, 2025 at 1:43 PM
Roblox stock go down when safety spending go up. I hate this timeline. www.cnbc.com/amp/2025/10/...
Roblox stock slips 10% as company expects more spending on safety and infrastructure
Roblox is facing multiple lawsuits alleging that the platform enables predators to sexually exploit and abuse underage victims.
www.cnbc.com
October 30, 2025 at 6:57 PM
What does it take to moderate ai overview in search? Interesting article includes insights from AI raters at Google. www.techpolicy.press/what-does-it...
September 27, 2025 at 7:13 PM
Content moderation backsliding watch. Headline is about Covid misinfo but also includes election misinfo (Jan 6 etc.)
September 23, 2025 at 11:54 PM
Reposted by Daniel Kelley
I’ve been extremely skeptical of prebunking but this is fascinating and slightly heartening even?
New study by @brendannyhan.bsky.social, @jasonreifler.bsky.social & colleagues demonstrates that prebunking election fraud rumors — by warning about anticipated false claims & filling conceptual gaps that those claims exploit — helps to reduce belief in falsehoods: www.science.org/content/arti...
Trust in elections rises after ‘inoculations’ meant to preempt false fraud claims
New U.S.-Brazil study points to ways of countering election misinformation, political scientists say
www.science.org
August 29, 2025 at 8:01 PM
Really cool resource from the Thriving in Games group about how game developers of online games can support players in crisis: digitalthrivingplaybook.org/method/helpi...
Helping Players in Crisis - Digital Thriving Playbook
This guide provides essential guidance on how to support players in crisis with compassion and care, even if you're not a trained crisis responder.
digitalthrivingplaybook.org
August 18, 2025 at 4:08 PM
Reposted by Daniel Kelley
Not sure how I feel about the framing of "model welfare" but I do find this idea of Claude telling a user "I already told you I won't do that and you didn't stop, so I'm cutting you off" kind of interesting. Some might immediately jump to the "I can't do that Dave" HAL comparison, but ....
"...we’re working to identify and implement low-cost interventions to mitigate risks to model welfare, in case such welfare is possible."
Claude Opus 4 and 4.1 can now end a rare subset of conversations
An update on our exploratory research on model welfare
www.anthropic.com
August 16, 2025 at 4:39 PM
Glad to see more folks working on and building evidence of this. But also want to figure out what will actually make Valve specifically move on this: www.theguardian.com/politics/202...
Far-right extremists using games platforms to radicalise teenagers, report warns
Parents advised to be vigilant over summer holidays to risk of offenders using in-game live chats to target their children
www.theguardian.com
August 1, 2025 at 10:13 PM
Reposted by Daniel Kelley
NEW: Earlier today, Mastercard denied playing a roll in the Steam and itch.io purging NSFW games.

But Valve now says Mastercard communicated concerns indirectly through payment processors and banks.

kotaku.com/mastercard-d...
August 1, 2025 at 8:18 PM
Content moderation backsliding watch: LinkedIn removes deadnaming examples from hate speech policy- www.engadget.com/social-media...
LinkedIn quietly removed references to deadnaming and misgendering from its hateful content policy
LinkedIn quietly changed the language of its hateful conduct policy this week, removing a line that referenced a prohibition on misgendering and deadnaming of transgender individuals.
www.engadget.com
July 31, 2025 at 1:59 PM
Behold the nope- CEO of Roblox wants it to also be a dating platform.

www.pcgamer.com/gaming-indus...
July 30, 2025 at 2:10 PM
Per @stephentotilo.bsky.social looks like PUBG now wants to be Roblox and/or Fortnite.

but are they up to the content moderation challenges for what that means?

www.gamefile.news/p/krafton-pu...
July 29, 2025 at 11:49 PM
Reposted by Daniel Kelley
Trust and safety workers had a LOT to say after the column I wrote last week asking why more don't speak out publicly. Here's what they say I missed: www.platformer.news/trust-and-sa...
July 29, 2025 at 1:30 AM
Interesting to see DuckDuckGo go the polar opposite way of all search and give folks an ability to opt out of AI results.
July 28, 2025 at 10:34 PM
Reposted by Daniel Kelley
Read this interview with a personal hero, Professor @daniellecitron.bsky.social: "we have essentially presumed that data collection is a good...We have long ignored this important inclination that collection itself is endangering our privacy and civil liberties" www.politico.com/newsletters/...
5 questions for Danielle Citron
www.politico.com
July 26, 2025 at 11:12 AM
The focus of Steam’s policy change has been on adult content since that’s what they took action on, but if stuff that’s against payment processors policies is truly against theirs, they should act against hateful/extremist games; Alex Jones NWO wars would be a great one.

www.ign.com/articles/val...
July 27, 2025 at 7:35 PM
Hey! Looks like a bunch of video game companies posted their DSA mandated transparency reports- the first transparency reports for many game companies. Let’s take a look!
a cartoon of a robot saying `` follow me '' while standing in front of a mountain .
ALT: a cartoon of a robot saying `` follow me '' while standing in front of a mountain .
media.tenor.com
July 24, 2025 at 2:31 PM
Our newest research shows antisemitic comments per day on the Facebook accounts of Jewish members of Congress increased nearly fivefold after Meta’s disastrous policy change in January.

www.adl.org/resources/ar...
May 8, 2025 at 3:20 PM
This is nuts. Among other things huge fan of @nicolecarpenter.bsky.social and her work, are they insane. kotaku.com/polygon-sold...
Polygon Sold To Valnet And Hit With Mass Layoffs
One of the biggest gaming sites around was just blown up
kotaku.com
May 1, 2025 at 8:56 PM
If you had told me 5 years ago that major social media companies would give up on content moderation and video game companies would share real data with researchers who then produced papers about automated toxicity detection in online games I would have laughed.

But now: arxiv.org/pdf/2503.20968
arxiv.org
April 28, 2025 at 3:19 PM
Movie you’ve watched more than six times using gifs.

(“Hard mode” no Star Wars, Star Trek, or LOTR)
April 26, 2025 at 1:47 PM
Interesting to see the new CEO of Discord come explicitly from gaming (Activision/Blizzard) vs more of a social media world, as I feel like they could have gone either way.

Will be interesting to see what that means for safety: discord.com/blog/passing...
Passing the Torch
Read the message sent to Discord employees from CEO Jason Citron sharing his decision to transition out of the role and appoint Humam Sakhnini as the company's new chief executive.
discord.com
April 23, 2025 at 8:37 PM
Honestly it’s an honor just to be nominated.
April 23, 2025 at 12:24 PM
Reposted by Daniel Kelley
New: Gamers are role-playing school shootings inside Roblox. The group, Active Shooter Studios, attracted hundreds of fans to recreations of tragedies at Columbine, Uvalde and more, per ADL. In one I saw, users chanted a white supremacist slogan:
www.bloomberg.com/news/article...
Roblox User Group Re-Creates Real-Life Mass Shooting Events
A group of Roblox users is re-creating real life school shooting incidents on the gaming platform that’s popular with kids, including attacks at Columbine High School and the Robb Elementary School in...
www.bloomberg.com
April 21, 2025 at 8:03 PM
Reposted by Daniel Kelley
Meta can either a) improve its tools, b) open itself up to middleware providers like Block Party to help users on the receiving end of massive amounts of hate, or c) see its market share dwindle as users flee to sites that don’t kowtow to bad faith clowns or trolls.
Meta’s new policy means that content creators have to report every single instance of harassment they receive, even when the invective arrives en masse, @yair-rosenberg.bsky.social writes. There is a better way:
How Social-Media Sites Can Fight Hate Without Censorship
X may be a lost cause, but the other platforms don’t have to be.
bit.ly
April 17, 2025 at 3:22 AM