The Autocrat's Algorithm

Is Google News helping to spread propaganda?

On Nov. 18, I used the enormously popular news aggregator Google News to search for information about the Russian alleged arms dealer Viktor Bout, recently extradited from Thailand to the United States. The top story was a dispatch from the state-controlled newswire RIA-Novosti, which essentially transcribed a statement from the Russian foreign minister demanding that Bout receive a fair trial. The other top results were a mixed bag, including Western sources like CBS News and Agence France-Presse as well as other Russian state-funded sources like ITAR-TASS and the television network Russia Today. (Typical U.S. headline: "Alleged 'Merchant of Death' Pleads Not Guilty." Typical Russian headline: "Bout was psychologically pressured during flight to U.S.")

Two weeks earlier, a search for "Myanmar Election" would have returned dispatches from U.S. sources like UPI and the Los Angeles Times, describing Burma's just-concluded poll as a highly rigged sham, but also an opinion piece from the Global Times, an internationally focused publication produced by China's People's Daily titled "Myanmar's Election a Step Forward."

Of course, offering news from different international perspectives is the whole point of Google News. The service was developed by Google's Krishna Bharat shortly after the 9/11 attacks with the goal, as he later put it, of "helping people understand multiple points of view, and hence becoming wiser for it -- whether they agree with it or not." But those points of view are often coming from state-sponsored news sources in countries, like Russia and China, where independent journalists are either harassed and persecuted or outright banned. Could Google News's level playing field be enabling authoritarian regimes to more easily get out their message?

"The web gives us the possibility to reach an audience who cannot watch us on TV, and who are more used to getting news online," Margarita Simonyan, editor in chief of television network Russia Today, told me in an email. RT, as it's more commonly known, was founded five years ago, partly by RIA-Novosti, and is widely seen as an effort to improve Russia's image around the world, though it denies having a pro-Kremlin bias.

Nonetheless, others detect a strong pro-Russian slant in the network's coverage of international events. During the 2008 Georgia-Russia war, RT accused the Georgian forces of "genocide," but reportedly instructed its reporters not to report from ethnically Georgian villages that had been attacked by Russian troops. The network has also been criticized for giving airtime to fringe anti-American political figures and 9/11 conspiracy theorists. Some mainstream Washington analysts -- including this author, once -- do appear on the network's broadcasts, but sometimes find it difficult to get their views across.

"It's a little hard to go back on a TV show that has 'mechanical difficulties' every time you're supposed to speak," says Council on Foreign Relations Russia analyst Stephen Sestanovich.  

Judging by the results for Russia-related queries, however, RT's website seems to be succeeding in spite of its editorial slant.

Google doesn't disclose the complex algorithm by which it ranks search results, though that doesn't stop news outlets (including this one) from trying to figure it out. "Search engine optimization," or SEO, has become an obsession for media outlets looking to gain an edge on the competition in the new journalism landscape created by Google.

In an extreme example of this trend, some new online news outlets such as Associated Content and Demand Media generate content purely based on Google search queries rather than any sort of journalistic value, and newspapers are beginning to experiment with the formula.

Simonyan wouldn't speak on how her network seems to perform so well on Google News, saying, "Only Google can explain how it works."

According to Google spokesman Chris Gaither, some criteria include the "freshness" and "localness" of the story. The site also judges the reliability of different sources by a number of criteria, including the number of repeat visits from users.

One reason state-sponsored media often rank so high in response to specific queries might be that they're often the main source of original information from the countries they cover. Informal studies have observed that Google tends to prioritize original reporting over re-reported content. With either shrinking news budgets or government restrictions preventing Western news agencies from covering events in countries like Iran and Russia, that gives state-sponsored outlets a clear edge. A search for the latest news on Iranian President Mahmoud Ahmadinejad is likely to turn up so many stories from loyal state-sponsored outlets like PressTV and Fars News because they spend a lot more time covering him and have much better access.

But some analysts wonder about the unintended consequences of this preference. "If no one's covering the story but a news wire, a bunch of sources copying the newswire, and a state broadcaster who's basically there to refute the news wire, is Google News doing the right thing for us by prioritizing that state broadcaster?" asks Ethan Zuckerman, senior researcher at Harvard University's Berkman Center for Internet and Society and co-founder of the international blog aggregator Global Voices.

For example, are RT and RIA-Novosti really trustworthy sources on the brutal beating of reporter Oleg Kashin, who was nearly killed last month, likely for engaging in just the sort of journalism that these sources fail to provide? Their stories mention the frequent attacks on journalists in Russia, but fail to note, as the New York Times and AP did, that these recent crimes have nearly all gone unsolved by the authorities.

Or consider the Fars News's coverage of Ayatollah Ali Khamenei's recent visit to the holy city of Qom, which takes note of the "astonishingly warm welcome" he received in the city, proving "the nation's strong relationship and support for the Islamic establishment." In the foreign press, the visit was widely seen as an effort by the supreme leader to shore up support among Iran's increasingly critical clerical elite.

It's probably not possible, or desirable, to ask Google News to get in the business of blocking or otherwise filtering news from objectionable sources. Propaganda is often in the eye of the beholder. While few American readers object to getting their news from the BBC, partially funded by the British Foreign Office, many find the Qatari-funded satellite network Al Jazeera hopelessly biased. Should it be blocked along with the Chinese wire service Xinhua? What about stories from the U.S. taxpayer-funded Voice of America and Radio Free Europe?

Policing content this way would directly contradict the search engine's guiding philosophy, which Zuckerman says has traditionally been about making the "algorithmic choice, rather than the human choice." On the other hand, Google already labels some stories as blog posts, press releases, or paid content, rather than reported articles, but Gaither explained the company's reluctance to expand labeling further.

"One thing we have not done is to label anything according to any perceived political bias," he says. "We feel that's a slippery slope that we don't really want to go down."

But perhaps the site could provide more information to let readers make those judgment calls on their own. "I would appreciate it if Google made it easier to check on the source and find out what we know about them, see their other coverage and maybe a profile of them. That strikes me as something that would be a reasonable way to handle it," Zuckerman says.

Google News is just one example of how the Internet is slowly shifting the balance of power for information. During the Cold War, the United States and Britain pumped Voice of America and the BBC World Service into communist countries, providing an alternative to state media, while copies of Pravda were pretty hard to come by in the West. Today, outlets like RT and Iran's PressTV are reversing the trend, providing an alternative narrative to global events.

"RT offers [U.S. readers] stories and opinions that urge them to question more," Simonyan says. "That's why we receive a huge amount of feedback saying: 'Thank you for reporting this story, we would never find it [in the mainstream media].'"

And just because an outlet is state-sponsored doesn't necessarily mean it's not a useful source. 

"Russia Profile, which is a RIA-Novosti magazine [is] astonishingly good," says Sestanovich. "I hope the editor doesn't get in trouble if I say that you can learn a lot about bad things that are going on in Russia from it. The journalists who work for it take professional self-respect seriously and they run high-quality pieces on all kinds of topics." 

But according to Jeffrey Gedmin, president of U.S.-funded Radio Free Europe/Radio Liberty (RFE-RL), American readers aren't always aware of what they're getting from more loyal state-sponsored outlets.

"They have very high production values and mix in enough real news to make it not appear completely lackey and prejudiced, but mostly it's very much a Kremlin propaganda effort," he says. "Do people know that what they're watching is a wholly owned subsidiary of a hostile authoritarian government?"

While a majority of Americans now get at least some of their news online and 62 percent of online news readers say they use the Internet to find out about international events, it's still far from clear how aware U.S. readers are of the different types of news sources available from aggregators like Google.

Some see the rise of authoritarian state news outlets as a strategic issue. Walter Isaacson, president of the Broadcasting Board of Governors, the independent U.S. agency that oversees RFE/RL and Voice of America, has described the growing influence of RT and its ilk in stark terms, saying, "We can't allow ourselves to be out-communicated by our enemies."

Simonyan rejects the notion that RT's broadcasts are part of a propaganda effort. "Government funding does not mean being biased, just as being corporately funded does not automatically mean being independent," she says. "Government funded and corporate funded media have one thing in common -- they compete for an audience, and it is up to viewers to decide what to watch, and what not to watch."

Of course, giving users the choice to hear all sides of the conversation is exactly what Google News is supposed to be for and why it's so popular. But that level of choice often means that users are exposed to sources with varying levels of reliability. 

In Google's view, readers are expected to take far more responsibility for figuring out who exactly is giving them their news. According to Gaither, that's exactly the point: "We provide lots of different links so you can see how different sites from different geographies and political perspectives are covering the news of the day so you can make your own judgment on which sources you trust."

In other words: Reader beware.



Obama's Handcuffs

Was the Ahmed Ghailani verdict a victory for the rule of law, or the final nail in the coffin of the Obama administration's attempt to try terrorists in civilian courts?

A U.S. federal criminal court in Manhattan made history Wednesday when it issued a verdict in the case of Ahmed Ghailani, the first former Guantánamo Bay detainee to be tried in an American civilian court. The result wasn't exactly what the White House was hoping for: Ghailani was acquitted on 284 out of 285 charges related to the 1998 U.S. embassy bombings in East Africa. He was found guilty of one count of conspiracy, which carries a possible life sentence, and a minimum sentence of 20 years in prison.

Both the Obama administration and its critics were quick to use the ruling as ammunition in the ongoing debate over how to balance national security and the rule of law in the war on terror. Rep. Peter King (R-NY) decried the verdict as a "total miscarriage of justice," and said that it exposes "the absolute insanity of the Obama administration's decision to try al Qaeda terrorists in civilian courts."

White House spokesman Robert Gibbs soon fired back. The trial "incapacitated somebody that has committed a terrorist act and because of that incapacitation is not going to threaten American lives," he said.  He also affirmed the president's continued commitment to closing Gitmo.

Who's right? To get some answers, Foreign Policy asked six legal experts, former policymakers, and human rights activists to tell us what they believe this verdict means. Here's what they had to say.

Tom Malinowski:

Cynical commentators are portraying the Ghailani verdict as a blow to civilian trials.  But in case facts still matter, we should remember: A civilian American jury convicted Ghailani, swiftly and with finality. He will serve at least 20 years.

Had he been tried before a military commission, the judges would have likely also thrown out evidence tainted by Bush-era torture. All of the members of Congress criticizing this verdict (including New York Rep. Peter King) voted for the reformed military commission rules that prohibited evidence derived from torture and cruelty and should know better.

If the U.S. government had sought instead to keep Ghailani in Guantánamo without charge, there is absolutely no certainty that it could have held him for as long as the 20 years -- minimum -- that he will serve under this sentence. Do the proponents of this option really think that U.S. courts will allow detainees to be held forever without trial, especially if, at some point in the next few years, Osama bin Laden and other top al Qaeda leaders are killed or captured, and the "war" against them is seen to be over? Under such an approach, the United States would not only be creating, for the first time in its history, a system of preventive national security detention, but the danger of releasing potentially dangerous people would be greater than if they were prosecuted in the civilian system.

President Obama now faces a decision about terrorism trials to come. In his National Archives speech last year, he laid out his policy -- that detainees who have violated American criminal laws and who can be prosecuted will be tried in civilian courts.

The president can do what he has said he believes is best for the country. Or he can allow local, not-in-my-backyard politics to influence what is fundamentally a decision about national security and the administration of justice. He can, in effect, create a new category of detainee:  those who can be prosecuted, but won't be for political reasons. Such a decision would not only undermine the rule of law, but diminish the president's authority and invite even more irresponsible political attacks on his counterterrorism policies. Fortunately, it's his choice.

Tom Malinowski is Washington advocacy director for Human Rights Watch.

Juan Zarate:

This verdict is hugely problematic for the United States in two fundamental ways. First, it demonstrates the difficulties of bringing high-level al Qaeda figures out of Guantánamo and injecting them in to the legal system. The questions of evidence, duration of detention, and the complications of cases that are old or are growing stale demonstrate that there are mechanical and substantive problems to bringing these types of cases to civilian courts.

More importantly, however, it reveals a fundamental tension between the reality that these are individuals the U.S. government will not release and the desire to hold them accountable in a criminal context in which their innocence is presumed by law. The fact that failure is an option in the criminal legal system -- that acquittals are possible -- raises the specter of a case in which an al Qaeda figure can be acquitted but not be released. Such an outcome will do fundamental damage to the central premise of criminal legal proceedings.

It's time for a step back -- for a principled legal approach that admits that the United States is at war with al Qaeda, provides a legal framework that allows for preventive detention, and permits the use of the criminal and military commission systems to hold terrorists accountable in certain, specific cases -- as with American citizens. The United States needs clarity and legitimacy in its long-term legal and policy approach to detaining terrorists in this long war. We are further away today than ever before from that clarity.

Juan Zarate was the first assistant secretary of the U.S. Treasury for terrorist financing and financial crimes, deputy national security advisor for counterterrorism (2005 to 2009), and is now a senior advisor at the Center for Strategic and International Studies.

Aziz Huq:

If there is any "lesson" to be learned from the Ghailani trial and its aftermath, it is that the politics of terrorism have scant connection to the law and policy of terrorism.

Perhaps the fairest way of reading the verdict -- especially given that at least one jurors is known to have been a hold-out in favor of outright acquittal -- is as a compromise. The jury split the baby, issuing a guilty verdict on one token conspiracy charge while acquitting on charges based on the identical acts.

This illustrates what the framers of the U.S. Constitution thought to be the central function of the jury: that a group of ordinary citizens could, as leading constitutional scholar Akhil Amar writes, exercise "de facto power" to acquit "against the evidence." The Ghailani verdict, in other words, may be a paradigmatic example of the jury in its original constitutional function of being a check on government action.

As for the politics, the Ghailani verdict will be used as another stick with which to beat the hang-dog Justice Department of Attorney General Eric Holder. Of course, those who take this position do not, and will not, explain why the case is a failure. They simply assume that Ghailani must be guilty and that a sentence of 20-plus years is not enough. They hardly pause to ask why a jury of ordinary New Yorkers -- people who live and work in and around the fallen Twin Towers -- would have acquitted him on most of the charges.

They do not dwell on the reasons for acquittal -- for example, the government's use of coercive means to extract evidence and its decision to detain him for years at Guantánamo while witnesses against him passed away or disappeared. Nor do they question whether the policies they endorse will ameliorate or worsen the problem they claim to decry. Least of all will they say what they really mean -- that the framers of the Constitution got it wrong when they assigned to ordinary juries the task of adjudicating guilt or innocence.

Aziz Huq is assistant professor of law at the University of Chicago.

Andrew Borene:

From the perspective of due process, the Ghailani verdict reinforces the argument that terrorism jury trials can be held successfully in civilian courts. That said, the case won't answer existing questions related to national security -- of which the use of federal criminal courts versus military commissions for Guantánamo detainees is just one.

Broader questions remain about the policy framework the United States will use to handle the challenge posed by terrorism in the years ahead. Is America at war with al Qaeda, its formally trained adherents, and any self-initiated followers? Is counterterrorism a law enforcement action, a military action, or some kind of hybrid? What are the accepted standards for reliable information deemed actionable for intelligence-gathering or targeting? How do those standards differ from reliable information deemed actionable by law enforcement and later admissible in any court or commission?

The public energy surrounding the Ghailani case could potentially spur positive movement in answering these questions. As a lawyer, I am hopeful that Congress will act to create conscientious, comprehensive detention legislation setting out standards for different categories of criminal defendants, combatants, and terrorists or war criminals. As an American citizen and veteran, I also fully expect our commander-in-chief to continue taking decisive, ethical action under the Authorization for the Use of Military Force resolution to capture or kill positively identified terrorists.

Andrew M. Borene is an attorney and was an adjunct lecturer at the University of Minnesota's Humphrey Institute. A former U.S. Marine officer and government lawyer, he is the editor of the American Bar Association's U.S. Intelligence Community Law Sourcebook.

Karen Greenberg:

The lesson I take from the Ghailani trial is that civilian trials can work, that it's possible to have a trial in which members of a New York jury -- who know what terrorism is like and how dangerous and destructive it can be -- can see where the facts seem to lead to a verdict of guilty on one count and where the facts seem to lead to a verdict of not guilty on other counts. It shows that there are citizens of this country who can make judgements in a way that is free of politics and who understand how to think about facts and the law within a trial setting.

On the matter of the excluded witness, Hussein Abebe, I'm not sure how valuable his testimony would actually have been. He was judged not only to have been located as a result of the enhanced interrogation of Ghailani but to be in many respects a less than credible witness when he appeared here in New York at a pre-trial hearing.

That may have been because he felt threatened by the Tanzanian national police, who constituted a troublesome presence during this trial. But there is no way of telling how his testimony would have stood up in cross examination. Arguably, his contributions may not have proved credible to this jury, just like other evidence presented by the prosecution seemed somewhat tainted by factors in Tanzania. 

Karen Greenberg is executive director of the Center on Law and Security at the New York University School of Law and author of The Least Worst Place: Guantanamo's First 100 Days. Her ongoing commentary on the Ghailani trial has appeared regularly on

Robert Chesney:

It was inevitable that participants in the larger debate over terrorism and detention would make too much of the Ghailani trial's result. An across-the-board conviction would have been touted on the left as proof that civilian criminal prosecution should completely displace both military commissions and indefinite preventive detention (subject to habeas review). An acquittal would have been touted on the right as proof of the opposite.

Neither argument would have been persuasive. In any one case, the efficacy of prosecution is deeply contingent on the particular mix of evidence, legal issues, judge, and jury. Failure or success in one instance is no guarantee that the same result will be achieved in the next case.  The ambiguous nature of the Ghailani verdict -- a near acquittal, but a conviction on a single count that could still could open the door for the maximum sentence of life in prison -- leaves both poles in the debate trying to make lemonade from lemons. They are both doing their best to portray this as either a sufficient victory (though it does not look that way to many) or a complete defeat (which is just ridiculous).

Much of the post-verdict debate has centered on the question of whether obstacles the government encountered in this case could have been avoided had Ghailani been tried by military commission. While there has been much talk of evidence that could have been admitted in a commission that was excluded in the civilian setting, there has been very little convincing explanation of why this would have been so. At the same time, there has been insufficient attention to the unique problems that would have arisen had the case been tried by military commission -- most notably the certainty of protracted appeals over the propriety of a military commission's ability to prosecute a conspiracy charge. The U.S. national debate would be better served if it focused less on the question of the proper prosecution venue, and a lot more on the question of whether and when it is worthwhile to pursue any prosecution under circumstances in which the government has already decided it will hold the defendant indefinitely even if it loses the case.

Robert Chesney is the Charles I. Francis Professor in Law at the University of Texas, a distinguished scholar at the Robert S. Strauss Center for International Security and Law, and a nonresident senior fellow at the Brookings Institution. He blogs on national security-related legal topics at