It is Election Day today in the United States. Next week French Socialist militants will pick their Party official candidate. As a whole our Western society offers a perpetual ballet of ballot casting. However entertaining this can be, democracy is or ought to be like all professional art forms, based on hard work..
I wish for example the candidates would give more thought to the challenge of Internet. Not just to master how to set up a site and start up a blog. More along the lines of the recent initiative taken according to Steve Lohr's report by the Massachusetts Institute of Technology and the University of Southampton (1). Led by Tim Berners-Lee, this new program will study "Web science". May I dare suggest he makes reading "the Machine Stops" (2) a prerequisite?
I may have switched the direction of my gaze from the Middle Ages back to the future but I stick to my metaphor of a glass half-full, half-empty (see 10/24 fillip). Crediting the half which is full, one can rejoice that the threat of Internet towards privacy, eprivacy in my own words, has been identified, if not overcome. But I have come to the conclusion that the dangers caused by recommendation mechanisms run amok have not received the same level of attention. Past fillips (9/12, 9/26, 10/10 and 10/24) have attempted to fill up that empty half but there is a need for perseverance.
Indeed I have already reviewed Google, the greatest recommendation mechanism on earth (9/12 fillip). Election Day reminds us today that the weakness we spotted, the lack of transparency, the absence of accountability are far from being benign. As Tom Zeller Jr. tells us (*), liberals have manipulated Google recommendations, aka its page ranking, in order to smear Republican candidates. So-called Google bombs are nothing new, nor the depths to which candidates dive in order to discredit their opponents.
What dismays me is the explicit refusal to even acknowledge the issue. From Google, arrogance (we do no evil, it comes from the others) and ignorance ("the integrity of the search product remains intact"). As if highly transient results were not enough to move markets (3). From the perpetrator, the negation of democracy ("I do not believe the practice will actually deceive most Internet users"). As if an election was never decided by a few votes.
Do not be mistaken. Evil was not created by the Internet. Before the dawn of the Information Age, newspapers were the most influential recommendation mechanism. Their circulation gave them the same power, the same responsibilities and the same smugness as Google today. Perhaps inspired by Mark Twain's premature obituary printed by the New York Journal, the old joke has a publisher tell a reader in a similar situation: "Unfortunately our newspaper cannot publish a correction, for it is never wrong. But I am most willing to insert a suitable entry in tomorrow's birth notices."
A few months ago an editorial proudly proclaimed its resolution to maintain the relevance of the Financial Times in the future (4). Here is a challenge then. Last month it published an article about the three contenders to represent the Socialist Party in the coming French presidential election (**). In the US edition at least, the article came with the candidates' photographs. What mechanism can truly compensate Mr Dominique Strauss-Kahn for having received Mr Laurent Fabius's likeness and vice-versa? Was it a dirty trick by Ms Ségolène Royal's campaign to confuse her opponents? I did not even find the correction, assuming the Financial Times printed one. As for excuses, the previous paragraph but one will do, I am sure.
I am keenly aware that neither Google nor the Financial Times design their products with evil in mind. I am deeply grateful I can benefit from their recommendations. But success creates unintended responsibilities which cannot be ignored. Ask Zillow.com, a web site which provides free home valuations based on publicly available data. As reported by Damon Darlin (***), it stands accused by a coalition of community activist groups of undervaluing the homes in black and Latino neighborhoods.
In my 8/22 fillip I have proposed to tackle the issue of recommendation mechanisms by evaluating three factors:
Other words can and have been used to convey similar notions. Accountability for instance is achieved by combining responsibility and a protest process. To establish responsibility and ground protests in facts rather than opinions, transparency of the recommendation mechanism itself is required.
- protest process
Never mind the words, the task itself is not trivial for success addicts its victims to power and money. Power in turn does not want to share its secrets as demanded by transparency. And monetary returns would be decreased by the burden of responsibilities, which the winner perceives as liabilities.
I do not shrink from the task. After all isn't the purpose of ePrio's technology (see 6/20, 6/27 fillips) to provide an answer. I promise to speak one day about the recommendation mechanism it includes. But election candidates may want an alternative to ePrio's revolution (5/30 fillip). That is, if they can find time away from the twin chores of money raising and mud slinging.
Will the victors at the ballot be able to rise to the challenge and deliver (10/3 fillip) before The Machine Stops?
- (*).........A New Campaign Tactic: Manipulating Google Data, by Tom Zeller Jr. (New York Times) - October 26, 2006
.............Gaming the Search Engine, in a Political Season, by Tom Zeller Jr. (New York Times) - November 6, 2006
- (**)......Royal support dwarfs Socialist rivals, by Martin Arnold (Financial Times) - October 4, 2006
- (***)....A Home Valuation Web Site Is Accused of Discrimination, by Damon Darlin (New York Times) - October 31, 2006
- (1) Group of University Researchers to Make Web Science a Field of Study, by Steve Lohr (New York Times) - November 2006
- (2) The Machine Stops, a short story by E.M. Forster also available online courtesy of Paul Rajlich
- (3) Spam works: Evidence from Stock touts and Corresponding Market Activity, by Laura Frieder and Jonathan Zittrain - July 2006
- (4) No comment, no FT, editorial column (Financial Times) - May 2006