« Managing a Library's Electronic Collection: The Long March to an Integrated Discovery and Access Platform | Main | An Illustration of BLaw-BNA Content Integration: Does the Petition Clause guarantee a right of reasonable access to the government for a redress of grievances? »
April 16, 2012
Google's Latest CAPTCHA Codes Draws Criticism
Here’s a little bit more from the technology front. Google has been accused of another privacy violation. Any tech company that owns a user base numbering in the billions can make a move that draws criticism. Sometimes that criticism is valid, and sometimes not. Take, for example, Google’s latest twist on CAPTCHA technology. We’re used to the string of characters in twisted fonts and colors as a way to distinguish between humans and machines. Google added words sourced from scanned books as a way of improving the text presentation of those titles. Now, the latest twist is to use address numbers drawn from pictures of houses snapped in creating the street view in Google Maps.
Big Brother Watch is not very happy about this development, calling it a serious privacy issue by identifying the individual number of people’s homes. Google is accused of furthering its own interests (gasp!). Google responds that there is no identifying information that ties a picture to an individual, particularly as the pictures are close cropped to the number in question. Techcrunch has some examples here. I would be surprised if anyone can identify the geographic locations of the sourced numbers from the presentation. Yes, Google knows where they come from, but they are not telling. They use the crowd-sourced information to create a more accurate version of Google Maps. I think we are well past the point where Street View is in its generic form a privacy violation.
One comment (quoted in the Telegraph) from Nick Pickles, Big Brother Watch director of privacy and civil liberties is "The 'Don't be evil' mantra appears to have been replaced with a thirst for knowledge." I hardly ever conflate collecting public information with evil. There is no doubt that some of the things Google has done with Street View are unsettling. The “accidental” collection of unprotected Wi-Fi data by some of the collection vehicles is one, as is initially not blurring individuals captured in embarrassing situations. The law, if not common sense, has prevailed in those situations. Using cropped images of house numbers or street signs as CAPTCHA codes is not one of those things that would make me scared of Google. [MG]
Many of the TechCrunch article commenters are on the same track - it seems likely that Google is using the service to crowdsource desciphering addresses that their automated scanning can't make out. At any moment they likely have thousands of people with a reCAPTCHA image. Give them all the same number to read, the significant majority response gets tagged as the correct answer, which gets assigned to the location the image was shot. It doesn't seem like a serious invasion to me, either, but a clever way of getting a lot of manual labor done for free. They probably do a similar thing with the blurriest of the text we see, filling out an illegible spot in Google Books etc.
Posted by: Kreig Kitts | Apr 18, 2012 12:21:45 PM