1. Forum
  2. USENET
  3. comp.os.linux.misc

  • Re: AI Can Spot Spoofed Sites With 99% Accuracy

    From c186282@c186282@nnada.net to comp.os.linux.misc on Fri Sep 18 23:45:31 2026
    From Newsgroup: comp.os.linux.misc

    On 9/18/26 05:05, Richard Kettlewell wrote:
    gazelle@shell.xmission.com (Kenny McCormack) writes:
    c186282 <c186282@nnada.net> wrote:
    ...
    Sorry ... I just smell ROT here. The utility of

    What is "ROT" ?

    https://en.wiktionary.org/wiki/rot#Noun sense 2, metaphorically (or
    possibly sense 4). Speedy likes to capitalize words for emphasis.

    CORRECT ! :-)

    How do you inject the cadence and rhythm of
    natural speech using only text ? Caps and
    "*"s and "-"s and such are a way.

    As for "rot" ... (note 'slight pause' trick)
    the reference is to the overall safety on
    online activities, especially business and
    banking, clearly decreasing sharply over the
    past decade.

    This AI trick is yet another attempted, but
    probably futile and privacy-compromising,
    little fix. Very soon someone will use an AI
    to cheat the 'protecting' AI.

    I have run into more sites lately offering
    "biometric" authentication. Fingerprints,
    face scans, voice matching. The problem is
    that they have to *store* (see emphasis trick)
    this detailed info somewhere and that seems
    likely to be insecure. Even the largest tech
    firms seem unable to prevent penetration. Then,
    kind of literally, they "have your face" and can
    mis-use it widely.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Andy Burns@usenet@andyburns.uk to comp.os.linux.misc on Sat Sep 19 12:39:38 2026
    From Newsgroup: comp.os.linux.misc

    c186282 wrote:

    https://techxplore.com/news/2026-09-ai-spoofed-website-accuracy.html

    A machine-learning system can identify fake, or spoofed,
    website addresses
    And by this time next week, a different AI overlord will have tuned
    *their* latest model to generate undetectable fake websites ...

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From c186282@c186282@nnada.net to comp.os.linux.misc on Sat Sep 19 23:17:53 2026
    From Newsgroup: comp.os.linux.misc

    On 9/19/26 07:39, Andy Burns wrote:
    c186282 wrote:

    https://techxplore.com/news/2026-09-ai-spoofed-website-accuracy.html

    A machine-learning system can identify fake, or spoofed,
    website addresses

    And by this time next week, a different AI overlord will have tuned
    *their* latest model to generate undetectable fake websites ...

    Spot on !

    As I implied elsewhere, the entire e-everything
    paradigm is breaking down FAST. Everything is now
    TOO easily spoofed/redirected. Even the big tech
    corps cannot keep yer data secure - China has
    proven it can cut right in. "AI" is now making
    all this 1000 times easier/quicker.

    Back to the Sears Catalog ? Mail-Order ? CASH ?

    Maybe yea.

    And no, we have to decide/transition rather
    quickly - no long time to ponder. The global
    cyber-war is ON, big time.

    Alas I'm afraid the industrial/infrastructure
    people will be TOO SLOW. Hope you don't live
    near a big chemical plant or refinery or nuke
    power facility .........

    --- Synchronet 3.22a-Linux NewsLink 1.2