Spam-Bots continue to evolve

[alexxkay]

I just got a piece of LJ-Spam that included what appears to be an automatically-generated "lolcat". The content of the source picture appeared to be random, and the text overlaid on it was a mix of the target payload and what appeared to be gibberish. But the font and the placement of the text were classic lolcat. The body of the post contains good English sentences that each make sense individually in an advertisement context. Read as a body, it gives a distinct whiff of a Chinese Room program, but one more advanced than any I've seen before. It could almost have been written by a human who wasn't a very good writer, and had some odd neurological quirks, though that's not how I'm betting.

The once-funny idea that we would see true AI achieved as a side-effect of the Spam Wars is seeming more plausible by the year...

Comments

[harvey-rrit.livejournal.com]

Perhaps it's just a problem with the ridgeway armiphlage.

[metahacker.livejournal.com]

The Twitter bots are getting interesting/scary too. Lots of them, they spit out text snippets pulled from who knows where (looks like other tweets plus Gutenberg), and friend each other.

Then they favorite one of your tweets, and deliver the payload in their profile. Can't help but wondering if they're trying to build critical mass of friend-networking so they look legit before taking coordinated action.

[fabrisse.livejournal.com]

If AI evolves out of spam, no wonder it turns on its creator in so many movies.

[harvey-rrit.livejournal.com]

Friended on the spot-- I want to read what you post!