The Associated Press reports on privacy questions related to a Netflix contest that the company has canceled. I’ve written before about the ease with which anonymized data can be de-anonymized, allowing for identification of the individuals connected to the data:
DVD-by-mail service Netflix Inc. has canceled a sequel to a $1 million movie-recommendation contest, avoiding a potential courtroom drama over the privacy rights of its subscribers.
The retreat announced Friday settles a lawsuit alleging Netflix’s plans to release millions of movie-rental records that could have illegally exposed sensitive information about its subscribers’ tastes and lifestyles.
The Federal Trade Commission also had raised questions about the company’s ability to protect customers’ privacy, Netflix disclosed Friday. […]
Netflix intended to release the movie records without any names or other personal information attached to the data, but critics contended that the protections wouldn’t be enough to guarantee anonymity.
Those arguments were supported by two University of Texas researchers who said they were able to sift through data that Netflix released in its first movie-picking contest to identify certain people who rated movies.