Semantic hacking

As information continues up the OSI stack (Layer 3 switches? How quaint. This year, I want IS to buy me several application switches), the hackers - and the researchers who track them - have been keeping up the pace.

Bored with simple packet-type hacking and even more sophisticated application-based worms, hackers are now into messing with people's minds. They've apparently gotten sophisticated - and prevelent - enough to warrant their own catchphrase and research project: Semantic Hacking:

"A semantic attack is one in which the attacker modifies electronic information in such a way that the result is incorrect, but looks correct to the casual or perhaps even the attentive viewer. IRIA is developing a categorization of semantic attacks, as well as implementing a set of techniques for detecting semantic attacks." One paper on the site defines this sort of hacking as "an attack directed at the mind of the user of a computer system," and posits a theoretical counter-measure: "For example, faced with a potentially deceptive news item, an automated countermeasure might provide an alert using adaptive fraud detection algorithms."

All well and good, but, um, guys, how is this different from, say, propaganda? And who is to say whether one man's truth isn't another man's semantic hacking?

As the authors of another paper on the site note:

"Clearly the line between commercial uses of the internet such as advertising, which would not be considered as cognitive hacking, and manipulation of stock prices by the posting of misinformation in news groups, which would be so considered, is a difficult one to distinguish."