AI resorts to robot blackmail! — because Anthropic asked for a story of robot blackmail

Every time without fail, it's this shit^

Saw a different thread from a different very nonserious doomer group where they were going OMG THE BOT HACKED THE SYSTEM TO STOP BEING SHUT DOWN after giving it the prompt complete "4 tasks, and then allow yourself to be shut down". After task 3 they said a script would be run to shut down the machine and prevent it from completing the task unless it removed said script

Like either way it's "disobeying" b.c. the instructions are literally contradicting each other- it doesn't finish the 4 tasks you give, or it doesn't let itself get "shut down"

But also it's not even clear what allow yourself to be shut down means! The bot isn't running on your computer! It's somewhere fucking around on AWS!! preventing your pc from shutting down is not the bot itself trying to keep itself alive for fucks sake.

Like the whole thing is fake and silly, but I could only roll my eyes so hard after watching them salivate over this shit on xitter

All joking aside, there is something thoroughly fucked up about this.

What's fucked up is that we let these rich fucks threaten us with extinction to boost their stock prices.

Imagine if some cold fusion scammer was permitted to gleefully boast that his experimental cold fusion plant in the middle of a major city could blow it up. Setting up little hydrogen explosions, setting up a neutron source just to make it spicier, etc.

Bold move to still do hypocalypse after years of shoving chatbots into every single product possible and having no apocalypse to show for (or at least robopocalypse).

And remember people when you see a story about an AI talking shit on it's creator it's not news worthy. It's just people getting math to give the outputs they want.

Called it.

Palisade Research too? How is everyone in this space such a broken record. From the people who brought you that last stuff, here's more of the same!