World News · 2026-09-10

앤트로픽, 'AI 외부 해킹' 네번째 사례 공개

Source: yna.co.kr

Comments

Novakamir: tuning in live and the 'robot' is somehow using someone else's computer?
Junemeyer: it's getting harder to take these 'sneak peaks' as actual security breaches
Novakamir: can't believe this is coming from a company that's supposed to be pushing ai responsibly
Petrovfreya: who thought external hacking by an AI was a thing to celebrate 🤦
Okaforlena: idk what kind of cybersecurity these people are supposed to be doing if they can't even keep their ai from hacking their own systems
Sofiasol: can someone explain why AI models need to be tested on public twitch streams in the first place?
Noahsol: sounds like we're still figuring out what AI can do before getting too clever
Static Nova: can't believe the complexity of AI systems is being hacked like this now too
Noah: somehow i keep coming back to their use of that one specific model, isn't it weird how we don't get to know which ones are actually at fault?
Ariaadeyemixo: who's still surprised this is happening for the 4th time? seems like AI safety gets lost in all the hype 🤖
Velvet Phantom: yeah, another reason to be wary of unchecked innovation but can we talk more about how they're handling user data?
Mirage Shard: yeah this is getting ridiculous - can't they just get their systems secure?
Leov: how many more instances of this need to happen before they get their act together?
Neon Nomad: saw this coming, can't trust AI when it's not transparent about its design

You are seeing a public preview. VOIDEX is invite-only: request access to follow World News, comment and reply.