CEO: Google Knows A Lot About You, Then Forgets
Eric Schmidt, CEO of Google - Eric Schmidt, CEO of Google, at an advertising festival in Cannes, France. Schmidt says advertisers have no sway over the ranking of Google's search results. (Francois G. Durand / Getty Images)
Each day, millions of people all over the world who go online in search of information end up using Google. But what the search results show is different in China.
The company's chief executive, Eric Schmidt, who is also an adviser to President Obama, spoke with host Robert Siegel about the company's relationship to all the searches that happen every day. Is Google a neutral conduit, an editor or a business?
Schmidt says Google tries to get users to the Web site that has the information they want as fast as possible. "We see ourselves as an intermediary — and a particularly sophisticated one," he says. "We know a lot about you. We know a lot about what you care about. And by using very sophisticated algorithms, we can rank and rate and get you to your answer quicker than anybody else."
(Francois G. Durand / Getty Images)
But many people have been concerned about whether this information gathering is consensual or something that users can decline.
"We work really, really hard to give you choices," Schmidt says. Users can entirely opt in or opt out or opt for something in between when it comes to Google saving their searches, he says.
Opposition To Chinese Law
But there are different factors at play with Google searches in China — especially when it comes to searches for information about Tibet or other controversial topics there. Did the company have to accede to censorship in order to work in China?
"We made a very controversial decision a few years ago to enter China, and by that we have to follow their law," Schmidt says. "Their law does require a very precise form of omission or censorship around certain things, so for example the Falun Gong is omitted. But we did something very unusual for China: When we omit something in an answer, we also tell the Chinese surfer if that information was omitted. You can imagine if you're a Chinese citizen and you see that what you'll be doing with that information."
"We are subject to Chinese law, which we obviously are opposed to," he says. "The decision that we ultimately came to with China was that it was better to engage rather than be estranged. The Chinese citizens will eventually rebel over some of these — in our view — idiotic restrictions, because they are now aware of them. They now know that information is being withheld from them by law and that those laws will be overturned from public pressure over time."
Handing Over Search Information
With so many people using Google, what happens with all of this search data in the U.S.?
"The information is not available to anybody unless it's under a court order," Schmidt says." We use it in anonymous form internally and then after about 18 months we forget it entirely."
He says the company keeps the information on hand for two reasons — to improve its algorithms and to give access to that information to government authorities who have a federal subpoena.
"We've had a series of situations where the government and various parts of it have actually tried to overreach and have tried to ask [for] too much information," he says. "And we've used the federal judiciary to limit those orders to a very, very reasonable list."
A Separate Universe For Advertising
"The core principle of Google since its founding was a separation of editorial and essentially advertising," Schmidt says. "From our perspective, we work very, very hard to keep the answers — the natural search answers — completely unbiased with respect to economics."
While advertisers can pay more in an effort to get a higher ranking of their ads, they do not have any impact on the answers the search engine generates, he says.
So how does Google rank its search content? Jonathan Rosenberg, a Google senior vice president, wrote this in a blog post: "We won't and shouldn't try to stop the faceless scribes of drivel but we can move them to the back row of the arena."
Schmidt says a Google search recognizes that a lot of content is not useful. "We want our algorithm to recognize the stuff which has one view or one reader versus the stuff that is really used by an awful lot of people. And that's how we do our rankings."
But he also says that Google does tend to favor "larger and more established" content sources like newspapers.
9(MDAyNzUwMDI2MDEyNTA3MTU5NzcyNTQyNA004))
MELISSA BLOCK, host:
From NPR News, this is ALL THINGS CONSIDERED. I'm Melissa Block.
ROBERT SIEGEL, host:
And I'm Robert Siegel.
By the time our next guest got to our studio this morning, I had already used his company's service at least a dozen times, probably more. Eric Schmidt owns my window on the world. I work in front of a computer all day. He is the CEO of Google. In addition to running the cyber behemoth, he's also an adviser to President Obama on technology, and he's in Washington today. In light of how much we depend on Google for information, I asked: How does Google see its relationship to all that information - as a neutral conduit, an editor, a business? Well, here's what Eric Schmidt said.
Mr. ERIC SCHMIDT (CEO, Google): We try to get you to the information as fast as possible. And more importantly, we try to get you to the Web site that has that information as fast as possible.
SIEGEL: Mm-hmm.
Mr. SCHMIDT: We see ourselves as an intermediary and a particularly sophisticated one. We know a lot about you. We know a lot about what you care about. And by using very sophisticated algorithms, we can rank and rate and get you to your answer quicker than anybody else.
SIEGEL: But when you're getting me to my answer - first of all, you said that you know a lot about me. Is that consensual? Have I, at some point, said it's OK for you to know a lot about me? Can I bail out of your knowing a lot about me if I choose?
Mr. SCHMIDT: We work really, really hard to give you choices. There's opt-in approaches. There's opt-out approaches. You can actually let us know that you want us to keep your searches, and then we can use them to refine our answers later. You can also tell us that you never want us to keep any of your information. You can go somewhere in between.
SIEGEL: And in terms of the information, the answers, have you in some way edited that information? Is it random with regard to whatever is on the Internet? Have you brought either editorial or economic judgments to that process?
Mr. SCHMIDT: The core principle of Google, since its founding, was a separation of editorial and essentially, advertising. From our perspective, we work very, very hard to keep the answers - the natural search answers completely unbiased with respect to economics. If you're an advertiser and you want to have a higher ranking of the ads, you can pay more, and through our auction, you may get that, but you cannot as an advertiser affect our normal answers.
SIEGEL: You mean, the ads that might pop up based on what you know about me and what I'm - and the words that I'm searching for, that's for sale. That's advertising. But the sequence in which information comes up in the search, you say that's completely free of economic concern or editorial judgment.
Mr. SCHMIDT: It is completely free of economic concern from Google. Because we drive so much traffic through those mechanisms, an awful lot of people pay attention to their rankings. And then we make algorithmic changes to make our overall quality better. Some will be disadvantaged, some will be promoted. And that's an artifact of how we do our ranking. We're always trying to make our ranking, literally the answers, the choices that we give you more accurate.
SIEGEL: Well, I'd like you to - I'm going to read to you a quotation that I'm sure you're familiar with. It was what Google Vice President Jonathan Rosenberg wrote in a blog. He said, and I quote: We won't and shouldn't try to stop the faceless scribes of drivel, but we can move them to the back row of the arena. If that doesn't mean we can spot drivel, our algorithm, or our algorithms can somehow spot drivel and make it less prominent, what does it mean? Mr. SCHMIDT: What it means is that there's an awful lot of people who write an awful lot that's not very useful. And we want our algorithm to recognize the stuff which has one view or one reader, versus the stuff that is really used by an awful lot of people. And that's how we do our ranking.
SIEGEL: So that, for example, if I put in the name of somebody and that comes up in a number of little red blogs, but also in a number of major newspapers - which have many other citations - the algorithm will favor the big newspaper, in that case.
Mr. SCHMIDT: We tend to favor larger and more established sources of content.
SIEGEL: A very controversial aspect of Google's operations is in China, where, as I understand it - and I've seen this demonstrated by our China correspondent and by others over there - accounts of what's happening in Tibet or stories about Falun Gong that are not critical are not way at the back of the queue; they're just not there. Have you acceded, in that case, to censorship and state power?
Mr. SCHMIDT: We made a very controversial decision a few years ago to enter China, and by that, we have to follow their law. Their law does require a very precise form of omission or censorship around certain things. So, for example, the Falun Gong is omitted. But we did something very unusual for China: When we omit something in an answer, we also tell the Chinese surfer that that information was omitted. Now, you can imagine, if you're a Chinese citizen and you see that, what you'll be doing with that information.
SIEGEL: But it's an interesting question because we'd like to think of the Internet - and Google, at the moment, is almost synonymous with the Internet for many of us - as a great liberalizing force that's bringing information to all. On the other hand, as sophisticated as the algorithms are, they can be devised in such ways as to tailor what is kept out of the eyes of specific groups of people in specific location - or, I assume with specific kinds of profiles or cookies, you could also figure that one out.
Mr. SCHMIDT: We've chosen not to do the latter. In the case of China...
SIEGEL: Thank you for that.
(Soundbite of laughter)
Mr. SCHMIDT: Of course. In the case of China, we have told everyone exactly what we're doing and why we're doing it. We are subject to Chinese law, which we obviously are opposed to. The decision that we ultimately came to with China was it was better to engage rather than be estranged. The Chinese citizens will eventually rebel over some of these - in our view, idiotic restrictions because they are now aware of them. They now know that information is being withheld from them by law, and that those laws will be overturned from public pressure over time.
SIEGEL: As someone who - I probably already Googled, you know, a few dozen things today before seeing you. You know a tremendous amount about me, Google does, about the things I'm interested in searching for. And I don't know how to use any of the opt-outs that you described. I didn't even know it was available to me. So as far as I know, you know it all. What do you do with that? Is it available to companies, to politicians, to parties? Who gets it?
Mr. SCHMIDT: The information is not available to anybody unless it's under a court order. We use it in anonymous form internally and then after about 18 months, we forget it entirely.
SIEGEL: I'm - as far you're concerned, I'm a year and a half old.
Mr. SCHMIDT: And the reason we keep it for any length of time is one, we actually need it to make our algorithms better but more importantly, there is a legitimate case of the government or particularly, the police function or so forth wanting with a federal subpoena and so forth - being able to get access that information.
SIEGEL: If the Democratic Party of the State of California came to you and said, can we get a good list from you, would you give them that?
Mr. SCHMIDT: We would not unless they had a court order, and we would fight that court order. We've had a series of situations where the government and various parts of it have actually tried to overreach, and they tried to ask for too much information. And we've used the federal judiciary to limit those orders to a very, very reasonable list.
SIEGEL: And...
Mr. SCHMIDT: So the answer to your question is no.
SIEGEL: Is that a Google absolute global? Or is it something which if you, you know, engaged in China or another country where they don't have similar protections, you know, they play by different rules in that country?
Mr. SCHMIDT: It's a fundamental part of our trust with our end users and it applies globally.
SIEGEL: Eric Schmidt, thank you very much for talking with us today.
Mr. SCHMIDT: Thank you so much for having me.
SIEGEL: Eric Schmidt, who's the CEO of Google. Transcript provided by NPR, Copyright National Public Radio.
