Post

Conversation

Comfortably Smug
@ComfortablySmug
Holy shit. Anthropic caught Iran using Claude to target Navy bases, automate dossiers on Americans, and run influence campaigns on US social media. Details are insane. The morons uploaded everything including the organization plans of the Ayatollah's funderal lmao
Image
Quote
Anthropic
@AnthropicAI
We're publishing our most detailed threat intelligence report to date. It covers how people tried to misuse Claude—for cyberattacks, influence operations, surveillance, biology, and building weapons—and how we found and stopped them. We disrupted every operation in the report,
389.1K
Views
View quotes
David Watson 🥑
Post your reply

st
@freemktsfreemen
This feels like the kind of thing we should not have made them aware of and then give them bad intel (see movie The Imitation Game)
851
Gill.i.am
@FailingTaoist
Seems made up. I do highly doubt after America and Israel started attacking Iran, Iranians were logging onto claude via highly monitored and disrupted internet access. Seems extremely low probability given Israel's involvement. Sure this wasn't Israel?
700
Jatin Garg
@jatingargiitk
The report has those as separate actors. The naval one is an Iran-nexus group that built targeting handbooks out of open sources. Ship transponder IDs, and a roster of US personnel scraped from captions on public military photos. The dossier work is a different case, and the
2.6K
CocaKova
@Jonnykov
Funny that's it's a reversal now😂😂 This has probably been going on for a loooonnnggg time haha crazy they're JUST NOW getting caught
2.5K
Jscott
@J_Scott_Art
Kinda dumb for them to announce that and now shit the tap off for that information source.
Image
1K
Snorri Páll
@snorripall
This is the equivalent of sending NSA copy of everything you are doing. They can't be that stupid can they?
92
J Douglas
@jdfink5
The 2nd sentence make ZERO sense. Please explain how they used Claude... is this a follow up Op from the Harry Potter guy?
71
AAE
@AAC0519
And we thought the problem was sharing companies' info with their competitors.
211
Raella Beth
@RaellaK
I can't wait for the new Ayatollah style to come out, said no American ever! Can't they train Claude to refuse the requests? Just say no to Terrorist mode guys! Not sure if this is impressive or sad?
11
Mick Jay
@mickjayofficial
Just sounds like more fear-mongering on Anthropic’s part. With their current psyop (Coxon), I don’t believe a word of their coms anymore.
2.6K
fizz
@fizzbhau
Okay. But it appears to have been pretty successful. The funeral went without any event, the navy bases are decimated, and Americans hate the war (no influence needed). Thanks Claude, finally an American company doing something good for the world
635
XAEN
@xainirl
leaving contingency plans for the supreme leader's funeral in an ai prompt log is next level statecraft
8
The AI Therapist
@TheAIShrink
Translated from French
Anthropic has turned Texas into servers. Claude no longer sleeps, she scans our tweets and thinks of you between two nighttime backpropagations.
368
Chance McClain
@TexasChance
They’re all like “look yall we aren’t effective altruist crazy-pants! We are patriots, too! (Eww but yeah) look how awesome we love America but btw our stuff is super powerful and needs oversight and not only cuz we are angling for an IPO but also cuz we are the frontier!”
20
Joel Cheney
@joelcheney
The theme seems to be that, all the sudden, AI is an existential threat. Weird. Just when open-source models were beginning to compete with the big dogs. Didn’t see this coming a mile away…
1.2K
knicksnihilist
@Sassupreme
these tools (if you can even call AGI a "tool") have become indispensable… controlling compute/intelligence/energy will be the only resources that actually matter soon
391
Mark Babasa
@Mark_Babasa
Iranian prompt: How do beat America? Claude: You don't have enough credits. Enter credit card information here ___
24