Alexios Mantzarlis
@mantzarlis.com
đ€ 3641
đ„ 1324
đ 2403
Tech, facts, and carbonara. mantzarlis.com
pinned post!
Today on
@indicator.media
: A handful of anonymous bots have taken over fact-checking on X. This isnât hyperbole. In the first three weeks of May, just eight AI contributors wrote 50.3% of all visible Community Notes on the platform.
loading . . .
8 AI bots now write 50% of Xâs Community Notes
The crowdsourced moderation project faces a choice between scale and legitimacy
https://indicator.media/p/8-ai-bots-now-write-50-of-x-s-community-notes
6 days ago
7
328
160
reposted by
Alexios Mantzarlis
Alex Mahadevan
about 7 hours ago
Inspired by the work of
@mantzarlis.com
at
@indicator.media
, I fired up my AI Community Notes bot. I hope it actually flags political misinformation. And I'll be testing lots of LLMs to see which one is best at it.
alexmahadevan.com/blog/communi...
loading . . .
Meet Kind Raspberry Chickadee, my AI fact-checker
My quest to build a bot that would actually fact-check politics on X's Community Notes.
https://alexmahadevan.com/blog/community-notes-bot-politics
1
2
2
Beautiful day to see Gotham FC â Jordyn Dudleyâs going places
1 day ago
0
4
0
I am, once again, tragically unrepresented at a World Cup. With Italy and Greece out, whom should I root for? Can't be a previous winner. Preferably a democracy. Currently leaning towards Canada (b/c Alphonso Davies, and also
@indicator.media
HQ)
1 day ago
1
1
0
reposted by
Alexios Mantzarlis
Jared Holt
3 days ago
I've seen researchers/reporters discount the importance of monitoring alternative platforms over the last few years, but *so many* conspiracy theories and false narratives spawn on those sites. This nonsense about hantavirus went around for three days before MAGA influencers started blasting it on X
add a skeleton here at some point
2
50
10
This is nuts
2 days ago
0
2
0
reposted by
Alexios Mantzarlis
AllyMcK
3 days ago
The lady opposite me on the train was reading this. To be fair she was only on the first chapter.
271
8166
989
reposted by
Alexios Mantzarlis
Philip Bump
3 days ago
This remains so funny to me. Note the âlast updatedâ line.
70
1574
188
reposted by
Alexios Mantzarlis
Cooper Lund
4 days ago
Hot take but: a plank you stand on should not be able to travel at interstate highway speeds
add a skeleton here at some point
13
398
45
hm, are those really the two options for children's titles
3 days ago
2
1
0
reposted by
Alexios Mantzarlis
Renee DiResta
4 days ago
No no surely this canât be true thereâs an anti-censorship EO I keep hearing about
www.engadget.com/2182752/doj-...
loading . . .
The DOJ is reportedly asking Reddit and X for the identities of anti-ICE posters - Engadget
Reddit and X have reportedly received grand jury subpoenas from the Justice Department, seeking the identities of users who posted about ICE.
https://www.engadget.com/2182752/doj-reportedly-asking-reddit-x-for-names-of-ice-critics/
5
65
24
Alt hed: Sham organization meant to undermine international institutions and make peace even harder struggles to get donors to honor their commitments
loading . . .
Donald Trumpâs Board of Peace fund is empty
Despite $17bn in pledges, organisation is stuck in limbo with no money flowing to projects in Gaza
https://www.ft.com/content/5ba3bd2c-0e0e-4306-84be-99d6d89b0d49?syn-25a6b1a6=1
4 days ago
0
1
0
reposted by
Alexios Mantzarlis
Celeste Labedz
6 days ago
"Surely we can all agree on the basic facts" doesn't always work, and the times it doesn't work are the most important times to get things right. If a platform's requirement for a fact-check is "will both sides find this palatable" rather than "is this true" then the plot is already lost.
2
10
3
reposted by
Alexios Mantzarlis
Celeste Labedz
6 days ago
I'm a former Community Notes super-user interviewed for this piece. Whether the notes are written by people or language-mimicking algorithms, the central flaw of Twitter's Community Notes is the same as it's always been: it's a "fact-checking" system that doesn't involve the checking of facts.
add a skeleton here at some point
3
24
11
Today on
@indicator.media
: A handful of anonymous bots have taken over fact-checking on X. This isnât hyperbole. In the first three weeks of May, just eight AI contributors wrote 50.3% of all visible Community Notes on the platform.
loading . . .
8 AI bots now write 50% of Xâs Community Notes
The crowdsourced moderation project faces a choice between scale and legitimacy
https://indicator.media/p/8-ai-bots-now-write-50-of-x-s-community-notes
6 days ago
7
328
160
lol (h/t
www.theverge.com/tech/936176/...
)
7 days ago
1
7
1
reposted by
Alexios Mantzarlis
Alex McMillan
9 days ago
YOUâRE LISTENING TO SISYPHUS FM, THE HOME OF NON-STOP ROCK
557
48946
15585
I'm working on a 3,000-word-long piece on Community Notes for real sickos only
10 days ago
0
5
1
"Itâs alienation on top of alienation. What value am I actually making by throwing this garbage into the world?"
loading . . .
The Filipino virtual assistants behind LinkedInâs âthought leadershipâ content mill
For $7 an hour, virtual assistants use AI tools to write LinkedIn posts and comments on behalf of Western executives.
https://restofworld.org/2026/virtual-assistant-linkedin-engagement/
11 days ago
1
0
0
Sam was writing about this stuff in 2017. Always read Sam about this stuff.
add a skeleton here at some point
11 days ago
0
6
1
reposted by
Alexios Mantzarlis
Craig Silverman
12 days ago
Absolutely awesome to see that my guide to using Skills in Chrome helped an investigator with his work! Here's the post:
indicator.media/p/the-indica...
add a skeleton here at some point
0
16
3
I both love and hate these businesses trying to game Google Maps searches with their names (h/t
@davidsullivan.bsky.social
who first alerted me to this phenomenon)
12 days ago
0
6
3
reposted by
Alexios Mantzarlis
Gordon Pennycook
17 days ago
There has been a LOT of research on misinformation over the last decade. In a new paper (for Annual Review of Psychology),
@dgrand.bsky.social
and I attempt to distill the key points of consensus, highlight major disagreements, and point to future directions.
papers.ssrn.com/sol3/papers....
loading . . .
<p>Consensus, Disagreements, and Open Questions in the Psychology of Misinformation</p>
Research on misinformation has expanded rapidly over the past decade. This review examines areas of consensus and contention within the literature on misinforma
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6752858
2
125
54
reposted by
Alexios Mantzarlis
Canadian Journalism Foundation
12 days ago
CJF Hinton Award Finalist Spotlight:
@indicator.media
co-founder
@mantzarlis.com
, with co-founder
@craigsilverman.bsky.social
, and co-authors Santiago Lakatos and Benjamin Shultz, was shortlisted for three articles from Indicatorâs body of reporting on AI nudifiers.
1
4
2
Starting today, tech platforms are required by the US Take It Down Act to remove synthetic non consensual intimate imagery. Here's what their removal workflows look like in practice:
loading . . .
Tech platforms are now required by US law to remove deepfake nudes. Hereâs how you can request a takedown.
Indicator reviewed the removal flows for 16 different platforms
https://indicator.media/p/tech-platforms-are-now-required-by-us-law-to-remove-deepfake-nudes-here-s-how-you-can-request-a-take
13 days ago
0
9
8
Iâm going to miss a whole lot of things about America â its morbid attachment to A/C isnât one of them
16 days ago
1
2
0
In May's Show & Tell,
@craigsilverman.bsky.social
and I are joined by Amnesty International's Barbara Marcolini. Barbara spoke about her role as a journalist in a human rights org. and walked us through how she documented the Israeli military's destruction of civilian infrastructure in Lebanon.
loading . . .
How an OSINT Investigator Documented the Israeli Armyâs Destruction of Southern Lebanon
YouTube video by Indicator
https://www.youtube.com/watch?v=9L7ibh-vFZ0
18 days ago
1
9
10
Marketers for pine nuts should note that âessential for pestoâ is more of a selling point than any of these attributes tbh
19 days ago
1
2
0
any chance anyone here is a riverside pro and can help me figure out why i can't get different layer to work sobs
19 days ago
1
0
0
reposted by
Alexios Mantzarlis
Julia Carrie Wong
19 days ago
can anyone recommend a person who can speak to the technical challenges/best practices for structuring/analyzing large data dumps? i'm not sure if i'm looking for an archivist or a librarian or someone else ...?
14
20
10
Hello -- I'm looking for active X Community Notes contributors who want to talk to me for a piece on
@indicator.media
. Happy to use their pseudonym in the story. Please share with your networks!
20 days ago
0
1
4
As a parent and as an expert of this awful beat, I 100% agree with this: "schools should remove pictures of pupilsâ faces from their websites and social media accounts because blackmailers are using them to create sexually explicit images, experts have said."
loading . . .
UK schools should remove pupilsâ online photos as AI blackmail threat grows, say experts
Criminals are manipulating pictures found on school websites and social media to create sexually explicit images
https://www.theguardian.com/technology/2026/may/08/uk-schools-remove-pupils-photos-online-ai-blackmail-threat-grows
21 days ago
0
24
9
reposted by
Alexios Mantzarlis
Judd Legum
24 days ago
Trump Media, the parent company of Truth Social, just reported its 2026 Q1 results. It lost $405 MILLION on $900,000 in revenue.
348
4200
1410
time to walk away from the computer
25 days ago
0
14
1
"at the heart of the issue is the way Meta has incentivised the creation of hateful AI slop as it chases user engagement and ad revenue"
loading . . .
The devout Muslim making a living from Islamophobic AI slop
A content creator in Pakistan is earning money from viral videos that stir up hate in UK
https://www.thebureauinvestigates.com/stories/2026-05-05/a-devout-muslim-in-pakistan-is-making-a-living-from-islamophobic-ai-slop
25 days ago
0
6
0
Lisa did not fold, did not seek a more palatable term than "misinformation", and did not apologize for doing the important work she does. Major respect â
add a skeleton here at some point
25 days ago
0
0
0
"Palantir said its revenue from U.S. commercial customers jumped 133% to $595 million, âwhile revenue from U.S. âgovernment customers rose 84% â to $687 million."
loading . . .
Palantir lifts annual revenue forecast on robust US government demand
Palantir Technologies raised its âannual revenue forecast and beat estimates for quarterly results on Monday, a sign of growing demand for its data analytics software from the U.S. government and comm...
https://www.reuters.com/business/palantir-tops-first-quarter-revenue-estimates-strong-us-government-demand-2026-05-04/
25 days ago
0
0
0
New on
@indicator.media
: At 11 a.m. on Wednesday morning, I discovered that my partner Mia has a hidden Tinder Platinum account. Except Mia is not actually my partner. She isnât even a real person. The image I uploaded was created with Google AI. So, it turns out, was the Tinder profile.
loading . . .
How "Cheater Buster" websites use fake info and dubious claims to trick people into paying
Indicator found an ecosystem rife with spurious claims, fake products, and AI-generated matches
https://indicator.media/p/how-cheater-buster-websites-use-fake-info-and-dubious-claims-to-trick-people-into-paying
25 days ago
2
62
34
> men's personal care products in the 2000s: "- For Men" after the brand name, in a blue container > men's personal care products today: "MANTASTIC MANTACULAR ALPHATRON (WITH ADDED PROTEIN)," vikings all over the packaging
25 days ago
1
7
2
oh boy I have a fun piece for tomorrow
26 days ago
0
5
0
reposted by
Alexios Mantzarlis
Kai Kupferschmidt
27 days ago
I've spent yesterday and today talking to researchers across the globe trying to understand the hantavirus outbreak aboard the cruise ship MV Hondius. My story in
@science.org
is here (and some threaded thoughts coming):
www.science.org/content/arti...
đ§Ș
#IDsky
loading . . .
Cruise shipâs hantavirus outbreak puts researchers in uncharted territory
Questions about the culprit virus and its route of spread remain as health officials make plans for stranded passengers
https://www.science.org/content/article/cruise-ship-s-hantavirus-outbreak-puts-researchers-uncharted-territory
21
817
504
reposted by
Alexios Mantzarlis
Eduardo SuĂĄrez
28 days ago
My colleague Gretel Kahn has spoken with 4 leading experts about how how AI, memes, trolling and slopaganda are reshaping the public sphere in light of the current conflict A timely piece. Thanks
@mantzarlis.com
@mahsaalimardani.bsky.social
reutersinstitute.politics.ox.ac.uk/news/trollin...
loading . . .
Trolling, memes and deepfakes: How AI is thickening the fog of war
The US and Iran are using technology to obstruct coverage of the current conflict. Four experts discuss how AI is reshaping war reporting and the news ecosystem.
https://reutersinstitute.politics.ox.ac.uk/news/trolling-memes-and-deepfakes-how-ai-thickening-fog-war
0
1
1
reposted by
Alexios Mantzarlis
Reuters Institute
28 days ago
NEW on our site How AI is thickening the fog of war With the US and Iran using AI and memes to obstruct conflict coverage, our own Gretel Kahn spoke to
@mantzarlis.com
@mahsaalimardani.bsky.social
Emerson Booking and Sam Dubberley from
@hrw.org
reutersinstitute.politics.ox.ac.uk/news/trollin...
loading . . .
Trolling, memes and deepfakes: How AI is thickening the fog of war
The US and Iran are using technology to obstruct coverage of the current conflict. Four experts discuss how AI is reshaping war reporting and the news ecosystem.
https://reutersinstitute.politics.ox.ac.uk/news/trolling-memes-and-deepfakes-how-ai-thickening-fog-war
1
18
12
reposted by
Alexios Mantzarlis
Dr. Pete Meyers
about 1 month ago
Look, before you mock people for having their database deleted in 9 seconds by Claude, you should know that deleting a database that thoroughly used to take up to 90 minutes.
20
983
223
bewildering stuff, I really hope these women get justice
loading . . .
These Men Allegedly Profit Off Teaching People How to Make AI Porn
Three Arizona women have filed a lawsuit against a group of men that alleges they used the womenâs photos to make AI porn influencers, then offered online courses showing others how to do it.
https://www.wired.com/story/ai-porn-lawsuit-arizona/
about 1 month ago
0
1
0
reposted by
Alexios Mantzarlis
jasmine mithani
about 1 month ago
Big news: The Minnesota legislature passed the nation's first ban on nudification apps, aka the generative AI apps behind the rise of nonconsensual deepfakes. The bill also allows survivors to sue for damages. Now the it heads to Gov. Tim Walz's desk.
19thnews.org/2026/04/minn...
loading . . .
Minnesota passes the nationâs first ban on ânudificationâ apps
The apps are one of the major ways nonconsensual AI deepfakes can be made without any technical expertise â including by kids.
https://19thnews.org/2026/04/minnesota-nudification-ban-ai-deepfake/
5
215
93
If you wanna scare an Italian friend (who doesnât live in the US) show them this price tag
about 1 month ago
1
1
0
reposted by
Alexios Mantzarlis
Chelsea J. Carter
about 1 month ago
Take a moment and read the
@indicator.media
report. It found 192 TikTok videos for
#hoax
#isolationchallenges
with 54.8 million cumulative views.
#factcheck
#tiktokchallenge
add a skeleton here at some point
0
1
3
Today on
@indicator.media
: Would you be willing to live alone in a remote Norwegian cabin for 90 days in exchange for $8 million? What about forgoing your phone to stay in a tropical overwater villa for $3.7 million?
loading . . .
How âisolation challengeâ videos on TikTok act as lures for online scams
TikTok removed hundreds of videos pushing hoax reward schemes after Indicator reached out
https://indicator.media/p/how-isolation-challenge-videos-on-tiktok-act-as-lures-for-online-scams
about 1 month ago
1
8
4
reposted by
Alexios Mantzarlis
David Ho
about 1 month ago
About 20 years ago, lighting was the single largest use of electricity in American homes. But energy-efficient lightbulbs has shrunk the category to sixth place!
www.nytimes.com/interactive/...
20
484
140
Load more
feeds!
log in