mastodon.world is one of the many independent Mastodon servers you can use to participate in the fediverse.
Generic Mastodon server for anyone to use.

Server stats:

9.4K
active users

#aisecurity

11 posts10 participants0 posts today

Whoa! 🤯 "Influence-as-a-Service"?! Are we seriously talking about AI pulling the strings for social media bots now? That's creeping into some next-level cyber territory, and honestly, it's unnerving. 😱

Look, as a pentester, I absolutely see how AI can be a game-changer for security – I get the upside. But *this* particular development? It’s something else entirely. Attackers are actively weaponizing this technology, leveraging it to pump out disinformation and manipulate how people think. And the kicker? It’s proving effective! 😩

It’s the same old story, isn't it? The "good" tools inevitably get twisted and misused by those with bad intentions. A classic scenario.

What we desperately need are sharper detection methods, plus a serious upgrade in media literacy across the board. How are you seeing this unfold? Are we stumbling towards a reality increasingly steered by AI? Let me know your thoughts. 🤖💬

#RSAc #AI #AIsecurity #standards

OASIS WEDNESDAY RSA RECAP: In today's #COSAI panel, OpenAI's Ian Brelinsky and @TrendMicro's Josiah Hagen led a technical session on Safe AI Adoption: Standards and Best Practices from coalitionforsecureai.org/
path.rsaconference.com/flow/rs

The @OASISopen and CoSAI team also can be found Thursday at our exhibit hall booth (South Hall #2256) during show floor hours, 10am - 2pm.

Coalition for Secure AIHome - Coalition for Secure AI
Replied in thread

@LukaszOlejnik

The assertion is, that one can not guarantee safety of #AI #LLM Models at all.

I'd like to get to the bottom of this, not that I doubt it by #AIsecurity is going to be increasingly more important, especially now that Old school #infosec has proven to be susceptible to the "Walk in and Seize control" exploit 😉

Edit;

Sometimes my friends tell me; Wulfy; /no_think

😁

#KIMissbrauch
Wenn #Sprachmodelle plötzlich menschenverachtende Aussagen treffen oder gefährliche Tipps geben, läuft etwas gewaltig schief.

Eine aktuelle Untersuchung zeigt, wie schnell #Feintuning unerwartete Folgen haben kann – mit brisanten Konsequenzen für reale Anwendungen.

Das Phänomen wurde nur durch Zufall entdeckt.

#KI #Sprachmodelle #Finetuning #AIRisiko #LLM #AISecurity

tino-eberl.de/missbrauch-kuens

Tino Eberl · Gefährliches Finetuning: KI-Modelle können aus dem Ruder laufen
More from Tino Eberl

Assuming for a moment the claims about #AGI and #ASI are even half true, there must be a rapidly evolving art of #AISecurity...

...which will prove as effective, as 30 years of #infosec proved effective
Against Elon Musks #DOGE perpetuating the largest theft of data in human history.

Asserting AGI and ASI will never happen because it's a scam.
Is no more responsible than a pointy haired boss not spending resources on security because hackers will ignore his firm.

ASI will give humans only a single opportunity to be right or wrong.
And they who assert ASI will never happen, are simply irresponsible.