U.S. Must Act Quickly to Avoid Risks From AI, Report Says - eviltoast
  • Syn_Attck@lemmy.today
    link
    fedilink
    English
    arrow-up
    3
    ·
    edit-2
    7 months ago

    The US government works with companies like OpenAI, they sell them custom capabilities while learning more about the system and then eventually they build and improve their own systems for more sensitive operational abilities. DoD has been heavily involved in AI (specifically generative AI) publicly for years now so you can be assured it’s been in research for at least a decsde… Here’s there github repo but you can find a lot of .mil PDF files talking about it too. https://github.com/deptofdefense/LLMs-at-DoD

    DARPA and the National Labs (Lawrence Livermore National Lab, Oak Ridge, Brookhaven, etc. are the ones who pull scientists from universities to work there when their research is something that has a clear and present military application.

    Also for the other guy talking about China not being good at AI or anything besides manufacture, here’s some good examples of their GenAI propaganda at work. Now, this isn’t actual AGI, which is what the article is worrying about, but it shows they’ve been working on these capabilities for some time and have an automated system for pushing messaging. https://m.youtube.com/watch?v=3FQSFnZpsqw

    I’m not sure if I’m worried more about AGI, or people I know, because agressive propaganda campaigns have been underway and are kicking up even more now, using LLMs and other GenAI.