But now, researchers have devised an attack that deciphers AI assistant responses with surprising accuracy. The technique exploits a side channel present in all of the major AI assistants, with the exception of Google Gemini. It then refines the fairly raw results through large language models specially trained for the task. The result: Someone with a passive adversary-in-the-middle position—meaning an adversary who can monitor the data packets passing between an AI assistant and the user—can infer the specific topic of 55 percent of all captured responses, usually with high word accuracy. The attack can deduce responses with perfect word accuracy 29 percent of the time.

  • ElephantInTheRoom@lemmy.ml
    link
    fedilink
    arrow-up
    5
    ·
    4 months ago

    Good thing I’m running my LLMs locally on a heavily encrypted PC with no network capabilities at all. Only way to not have my data siphoned, be it by hackers or big tech.

  • JackGreenEarth@lemm.ee
    link
    fedilink
    English
    arrow-up
    2
    ·
    4 months ago

    Does this affect locally run LLMs through summat like Jan? I already knew cloud based LLMs weren’t at all private, and thus don’t use them for anything I don’t care the public knowing about.