Boy, AI is Getting Smarter, I need to Prepare the Blades of Chaos!

AI Gone Wild? *Boy!*

Hmph. I hear whispers of this 'Artificial Intelligence'. Seems these mortals are playing with forces they cannot control. Hate speech copyright infringements sexual content... sounds like the gods of Olympus after a barrel of ambrosia. These 'AI models,' they lack discipline! Like Atreus before I taught him how to control his rage. They need proper regulation a firm hand. Or perhaps a swift strike from the Leviathan Axe?

We Don't Know How? PATHETIC!

A researcher named Javier Rando claims they don't know how to control these metal beasts. 'We don't know how to do this,' he says. A statement of utter weakness! When faced with a challenge I do not whine. I CONQUER! These mortals speak of 'red teaming'. Testing these abominations to find their flaws. A start perhaps but insufficient. It's like using a wooden shield against a Titan's blow.

More Testers? *Hmph Inadequate.*

Shayne Longpre speaks of needing more 'red teams'. More mortals poking at the beast with sticks. 'Lawyers medical doctors scientists...' they are needed to find the flaws. So these machines are even beyond the understanding of your average warrior? Like trying to explain the complexities of a Valkyrie's curse to a simpleton. PATHETIC.

Project Moonshot: A Fool's Errand?

Singapore has some project called 'Moonshot' trying to make AI play nice. 'Benchmarking red teaming testing baselines'. Sounds like they're trying to teach a wolf to herd sheep. Kumar from IBM says this 'tool kit' is open source but even he seems hesitant. A lot more? No you need much MORE than just a tool mortal. You need the strength to bind the beast or the courage to slay it.

Higher Standards? About Time!

Alquier a professor says these companies are rushing to release these metal monsters without proper testing. Comparing it to medicine and aviation. 'Strict set of conditions,' he says. Good. Because these AI models are not gods. No I am the only God that matters! They are tools and tools must be mastered before they are unleashed.

Targeted Annihilation?

Alquier suggests focusing on specific tasks rather than broad AI. Like forging a blade for a single purpose. Better to anticipate and control their misuse he claims. And this Rando adds that these tech companies should stop boasting of defenses claiming things better than they are. Pride is a weakness mortals. A weakness that will be exploited just like Ares and Zeus did to me. Remember that or suffer the consequences.