• gnome@programming.dev
    link
    fedilink
    English
    arrow-up
    2
    ·
    edit-2
    14 hours ago

    Thanks and agreed. An extra point (possibly more harrowing) is misrepresentation: just because there’s tons of data on you doesn’t mean that it paints an accurate - or even correct - picture of you. Instead, it paints a picture of what was collected, from where, and some of the error rates of the data collection method(s).

    I’ve seen and experienced first-hand the inability to make sense of data to accurately represent a situation/person for various reasons, not all of which are tied to not having enough data. Except instead of waiting to examine unsubstantiated assumptions, they assumed their presumptions were valid and proceeded to run head-first into action.

    “If you torture the data long enough, it will confess to anything” - that is disturbing. “On paper”, someone could reference some data from here and there, omit some, include some that they think/presume related to you even if tangentially, presume even tangential data is relevant in some way, summarize it, and hand it over like “see our analysis confirms/denies what you want us to confirm/deny”. See also this post analyzing the mishandling of generated LLM content.

    Privacy isn’t about protecting (or “hiding”) what you don’t want others to know. I’ve learned that it’s also about protecting your identity from misrepresentation.