The reds are coming! They're in your blog!
After posting the last item, the Blogger banner on this page was updated with a new set of "Related Searches" terms, which I gather are auto-generated from the content of the page, and pitched to encourage searches on Google for the same.
I am fascinated to observe the computed context for this blog includes "cnn news, australian news, cnn sports, news archives, art news, associated press, communism, cnn, news, king 5 news ".
One expects a certain amount of laissez-faire, which may explain how "King5 news" got there, or the "cnn sport", neither of which are actually relevant. However, the term "communism" is neatly slipped in there, a suggestive slant on the content of this site perhaps?
This editor is not now, nor has ever been, a member of the Communist Party;) Nor does the term occur anywhere within the current blog or its archives. One wonders what logic, if any, applies to the selection of "Related Search" terms; certainly a term like "communism" can be used to colour a reader's perception of a site.
Any observations on how "Pixelated Semantics" was taken over by bolsheviks without my knowledge are welcome.
Update: I'm beginning to enjoy this. On posting the current item, the Adsense banner was updated again to include terms "socialism" and "totalitarianism". One could continue in a recursive loop, commenting on the banner, and then observing the banner's changes in response to the comments, except it would probably amount to literary onanism.
I am now fascinated to find out how exactly these terms are generated. Unfortunately querying Google itself produces little information on the contexual algorithm. Some discussion threads on the Adsense feature here however are interesting.
We are creating a world in which contextual algorithms such as this make important decisions about ourselves. Technically speaking, they "carry out [...] analysis without receiving a hypothesis from the human analysts".
If such machines are capable of mis-interpreting the intention of this blog as "communist", then imagine if such "intelligence" was used to rate your trustworthiness for employment, suitability for loans, loyalty to government, or capability of criminality. Feel comfortable with that? Take a look at what banks , governments, and law enforcement agencies are doing with the technology. Developments include a "system which could scan crowds and pick out suspicious activity based on computer algorithms".
Item posted by AutoEditor at 12:10 pm ::
