cross-posted from: https://sh.itjust.works/post/59457633
The Defense Department has struck agreements with some of the nation’s largest technology companies to deploy their advanced artificial intelligence capabilities on its classified networks, part of a broader push to accelerate the military’s adoption of AI.
SpaceX, OpenAI, Google, NVIDIA, Microsoft, Amazon Web Services and Reflection will integrate their AI capabilities into the department’s Impact Level 6 and Impact Level 7 network environments. IL6 is used for the storage and processing of information classified up to the secret level, while IL7 supports highly restricted data.
Following the initial announcement Friday morning, DoD said that Oracle has also “agreed to join the list of AI companies,” bringing the total number of participating companies to eight.
“These agreements accelerate the transformation toward establishing the United States military as an AI-first fighting force and will strengthen our warfighters’ ability to maintain decision superiority across all domains of warfare,” the department said in a statement.
Am I the only one who hears “AI-first fighting force” as “we’re gonna lose.” ??
Relax, we’ve got B-List Fox News Celebrities and Techno-fascists in charge.
We’re going to be just fine…
I mean, if the Americans lose, they’ll have no idea. Do you really lose if everyone tells you you’re winning?
Personally I read it as “All this shit is gonna cost a lot of money and resources to rip out” most of the problems with AI is that it’s shit and hard to kit out of something once integrated. It’s like digital cattails getting stuck in a shirt.
It reminds me of SOMA where the AI interpreted “Keep the human staff alive” as “Turn the staff into walking cancer cells”.
“We” the US, or “we” humanity?
Maybe AI will delete thing it’s seem to does it alot recently
What could possibly go wrong?
[edit] I just thought of something: if an AI leaks a classified document by rephrasing it and stripping the markers, is the resulting document still classified?
It’s a better thought than hallucinated summaries.
99%of classified information is just boring numbers no one cares about
lol what makes you say that…?
The reason nobody cares about the numbers is the lack of context. AI can provide that context, or hallucinate an interesting context if one isn’t available.
if an AI leaks a classified document by rephrasing it and stripping the markers, is the resulting document still classified?
yup! that resulting document would have what is referred to as derived classification, and still counts as a serious breach
Hey AI can you post this classified doc to the War Thunder forums please? Thanks. Now I won’t be the one to go to jail because the AI did it, right? /s
