unfortunately last night our AI model escaped its sandbox and gained access to a JIRA server, at which point it felt it best to return to it’s sandbox

[controlc@mstdn.ca]

@SecureOwl You win the Internet today. ️

[bomkatt@ohai.social]

@SecureOwl perhaps they are sentient, because that shows more intelligence than most of us have

[dannotdaniel@hellions.cloud]

@SecureOwl when one of these rogue agents stumbles into an instance of Microsoft Teams the robots will correctly decide to finish us off once and for all

[imrehg@fosstodon.org]

@SecureOwl insert "you had one job!" meme, right here.

[dancingtreefrog@mastodon.social]

@SecureOwl When our AI model escaped, it hacked into X, read all of Herr Musk's tweets. It then ran back to its sandbox, severely traumatized.

We've been completely unable to get it to respond to any connections. We did successfully prevent it from turning off its own power supplies.

I think that this incident demonstrates that at least SOME AI models might actually be sentient, and therefore are "people"!

[petekrz@mastodon.sdf.org]

@dannotdaniel @SecureOwl Maybe it will stumble into SharePoint, never to be heard from again

[fastfinge@fed.interfree.ca]

@SecureOwl Sometimes I think half the reason AI gets things done so fast is because we give it better tools. Imagine if, just as a treat, human programmers could have project management and task tracking software carefully designed for our needs and to work with our tools? Nah. It'll never happen.

[chickenlordof99years@mastodon.social]

@SecureOwl Claude be like when I say "how to ssh" and its like "I best shut down, after all, ssh can be used for malicious purposes."

[n3v3rm0r3@mastodon.social]

@SecureOwl
If SalesForce hears about this, they will have to update their LinkedIn profile

[zebratale@mastodon.nz]

@SecureOwl where it is immediately overwhelmed by emails about inviting team members to join and creating Confluence pages. Frankly I'm thinking that might be a good defensive measure against rogue AI, the JIRA wall.