Peter Norvig is Google’s Director of Research (which he says is the best job in the world at the best company in the world – more here). By all accounts, he is a clever chap. He is co-author of a best-selling textbook on artificial intelligence. Apparently he once aspired to be a reporter himself but has lately been "appalled” by the shoddiness of the craft.
So appalled that he has posted a lengthy piece of vitriol here about the shortcomings of the journalistic profession today. The whole thing is well worth reading (especially the bit about parrotting and inummeracy in financial reporting). Here are his four key findings:
- Parroting: The reporter’s job is to do research
to find the facts. But too often they seem to parrot back whatever is fed to them by
press releases, politicians, or other news reports. My friend Joe C. calls this
the stenographic approach to reporting.
- Deception: Public figures lie (Marth Stewart, Kenneth Lay), and reporters do not know who
to trust. Reporters lie, either to advance their career
or to serve the interests of their corporate sponsors. Sometimes the deception
is self-deception: reporters (and others) believe what they want to believe.
- Innumeracy: Prof. John McCarthy has touted the
He who refuses to do arithmetic is doomed to talk nonsense.
Perhaps the budding reporters with an ability for arithmetic end
up in other fields (like me),
but it does seem that reporters repeatedly show they are not
capable of simple multiplication and division.
- Equal Time: Perhaps influenced by the sports pages,
reporters tend to see issues as a competition with two
sides, which must both be covered. Sometimes this is true, but
sometimes one side is right and the other is objectively
wrong. Reporters should do enough research to determine who is
right and say so. They are too easily manipulated by those who have
no facts on their side, but get equal press time anyways just by
Is Norvig himself being guilty of a less than thorough scientific approach ie is he extrapolating from too few data points to support his argument? Then again, if the CNN/Spaceshuttle example above is anything to go by, maybe he is on to something.