One evening in the fall of 2015, the writer Edward Jay Epstein arranged to have dinner at an Italian restaurant on the Upper East Side with the director Oliver Stone. At the time, Stone was completing Snowden, an admiring biopic about the former intelligence contractor Edward J. Snowden, who disclosed a vast trove of classified documents about National Security Agency surveillance programs to journalists in June 2013 and had since been living as a fugitive in Russia. Epstein was working on a book about the same topic, which has now been published under the title How America Lost Its Secrets: Edward Snowden, the Man and the Theft. As the writer recounts in that book, their conversation took a testy turn:

Toward the end of our dinner, Stone told me that he did not know I was writing a book about Snowden until a few weeks earlier. He learned of my book from Snowden himself. He said Snowden had expressed concern to him about the direction of the book I was writing. “What is it about?” Stone asked me.

I was taken aback. I had no idea that Snowden was aware of my book. (I had not tried to contact him.) I told Stone that I considered Snowden an extraordinary man who had changed history and was intentionally vague in my description of my book’s contents. Stone seemed to be reassured….

Edward Snowden
Edward Snowden; drawing by James Ferguson

Epstein and Stone had a history of rivalry when it came to interpreting another important historical event: the assassination of President John F. Kennedy. Early in his career, Epstein wrote three books about that topic. The first, Inquest (1966), poked holes in the rigor of the Warren Commission’s official investigation. The second, Counterplot (1969), brought a skeptical eye to the investigation by New Orleans District Attorney Jim Garrison, who pursued the theory that the Central Intelligence Agency orchestrated the president’s murder. And the third, Legend (1978), pointed readers to the conclusion that Oswald’s image as a mixed-up loner with half-baked Marxist ideas was an operational cover story—a “legend”—and that he had been a Soviet intelligence agent. (After the Soviet Union collapsed, the opening of the KGB’s archives did not corroborate the theory that Oswald had actually been a trained intelligence agent.)

Stone waded into those same murky waters with his 1991 movie JFK, which used a fictionalized version of Garrison’s investigation as a means to explore the theory that a right-wing conspiracy, spanning the CIA and the military-industrial complex, had been responsible for Kennedy’s death. The following year, Stone and Epstein were invited to be part of a panel discussion at New York’s Town Hall about the Kennedy assassination and the film’s controversial blending of fact and fiction. In preparation, according to a diary entry on Epstein’s website, he brought a 3×5 card on which he wrote:

Although they may aim at the same purpose of finding truth, non-fiction and fiction are two distinct forms of knowledge. The writer of non-fiction is limited by the universe of discoverable fact. He cannot make up what he does not know—no matter how strong his intuition or suspicion. The writer of fiction knows no such boundary: He can fill in whatever gaps exist with his imagination.

Now, years later, the two men once again found themselves eying each other as they circled the Snowden saga.

The conventional understanding of Snowden is that he was what he appeared to be: a computer worker in the intelligence world who became alarmed about the hidden growth of the American surveillance state and decided to reveal its operations to the world, copied archives of documents, and handed them to journalists whom he had summoned to Hong Kong and whom he entrusted to decide what to publish. Within the mainstream spectrum of interpretations of his actions, at one end are civil libertarians who consider him simply to be a heroic whistle-blower. At the other extreme are members of the national security establishment who consider him nothing more than a destructive traitor. In between are a range of those who think some of his disclosures met the high standard for “whistle-blowing”; that other disclosures brought to light important things that should not have been kept secret in a democracy—but that were also not necessarily, in and of themselves, abuses or overreaches; and that still other disclosures went too far and were not a public service.

Stone’s movie, which premiered in September, presents a comic-book version of the pro-Snowden narrative in which a wunderkind super-hacker takes on Big Brother. In telling that story, Stone mixes accurate material with fiction, while simplifying away complexities. His movie steps on the genuine privacy issues raised by Snowden’s disclosures with melodramatic embellishments, such as a scene in which an invented senior NSA official, his Orwellian face filling a floor-to-ceiling screen, casually reveals that he knows whether the Snowden character’s girlfriend is sleeping with another man. It omits actual Snowden disclosures whose individual privacy rationale was debatable, such as when he showed the South China Morning Post documents about the NSA’s hacking into certain institutional computers in China. And its discussion of the volume of Internet metadata the NSA collects from equipment inside the United States ignores any distinction between truly domestic e-mails and foreign-to-foreign messages that are merely traveling across domestic network switches.


Epstein’s book, by contrast, presents a negative view of Snowden. But the two works are not equivalent: Epstein does not merely oversimplify with the purpose of downplaying the benefits of Snowden’s leaks and emphasizing the harms. Rather he contends that the conventional narrative of what happened may have been a deceptive cover story. Epstein lays out the case that behind his image as a whistle-blower Snowden was instead an “espionage source” for Russia—perhaps its dupe at first, or perhaps its willing spy all along:

The counterintelligence issue was not if this US intelligence defector in Moscow was under Russian control but when he came under it. There were three possible time periods when Snowden might have been brought under control by the Russian intelligence service: while he was still working for the NSA; after he arrived in Hong Kong on May 20, 2013; or after he arrived in Russia on June 23, 2013.


The reader should know that Laura Poitras, one of the journalists to whom Snowden leaked documents in Hong Kong, later shared some of them with me, and we developed several articles from them for The New York Times. In addition, as part of a book on national security, I wrote a history of how surveillance technology, law, and policy secretly evolved in the decades following Congress’s enactment of the Foreign Intelligence Surveillance Act of 1978.1

It explained how the rise of fiber-optic networks in the late 1980s and the Internet in the 1990s placed mounting pressure on legal constraints written for the analog telephone era; how the Bush administration bypassed those rules after September 11 and then enlisted the Foreign Intelligence Surveillance Court and Congress to legalize what it had created lawlessly; and how the Obama administration decided to keep and entrench what it inherited.

I could not have written that history without the files disclosed by Snowden and information the government declassified because of his leaks. While there had been stray glimpses for years suggesting that the NSA was becoming far more powerful, facts were scarce and speculation and conspiracy theories had filled the void. Snowden’s disclosures enabled us to understand what was real about the NSA’s activities so we could engage in an informed public debate about the rules for twenty-first-century surveillance. This is why I regret Stone’s reintroduction of distortions into discussion of surveillance, and it may also color my reaction to Epstein’s book.

Snowden’s disclosures indeed prompted robust debate and policy changes. An appeals court ruled that the NSA’s bulk collection of domestic calling records was illegal, rejecting a dubious legal theory that the intelligence court had been secretly relying on for years. Congress ended that bulk collection program and required the intelligence court to tell the public when it issued novel and significant interpretations of surveillance laws. President Obama imposed unprecedented privacy protections for information about non-Americans that the NSA collects abroad. Technology giants like Google and many ordinary people began taking steps to more firmly secure their private information from hackers. Still, this enlightenment came at an undeniable, if difficult to measure, cost. Some terrorists, criminals, and unsavory regimes learned from Snowden, too, becoming harder to monitor and thereby making the world more dangerous.

Assessing whether Snowden’s disclosures served the public interest—whether they did more good than harm—turns in part on who counts as “the public.” Snowden’s critics, including Epstein, tend to define the public in nationalist terms, focusing their criticism on his disclosures about NSA operations abroad, where few domestic legal rules apply and the agency can indiscriminately vacuum up private messages in bulk. Snowden’s supporters point out that domestic data are also found abroad in the Internet era and they argue that consideration of the NSA’s work should take account of its effects on human rights: non-Americans have privacy rights, too.

Another complication for judging Snowden’s actions is that we do not know how many and which documents he took. Investigators determined only that he “touched” about 1.5 million files—essentially those that were indexed by a search program he used to trawl NSA servers. Many of those files are said to pertain to military and intelligence tools and activities that did not bear on the protection of individual privacy. Snowden’s skeptics assume that he stole every such file. His supporters assume that he did not. In any case they believe his statements that after giving certain NSA archives to the journalists in Hong Kong, he destroyed his hard drives and brought no files to Russia.


Epstein sees Snowden’s supporters as naive. He draws on his connections with the late James Jesus Angleton, the CIA’s paranoid hunter for KGB moles both real and imagined during the height of the cold war; after his dismissal from the agency in 1974, Angleton became an important source for Epstein, including for his book on Oswald. Much of How America Lost Its Secrets consists of Epstein building “alternative scenarios” like a counterintelligence investigator in Angleton’s mold trying to pierce presumed Russian deception. This, he concedes,

differs from that of a conventional forensic investigation aimed at finding pieces of evidence that can be used to persuade a jury in a courtroom…. The point is to assure that any alternative that fits the relevant facts, no matter how implausible it may initially seem to be, is not neglected.

And so Epstein asks: What if Snowden told secrets to Russian intelligence officials or brought files to Moscow, despite saying otherwise? What if he meant to end up in Russia all along, and it was just a cover story when he said he was trying to get to South America and was stranded in Moscow because the United States revoked his passport? What if Snowden sold out to China and/or Russia in Hong Kong? What if the Russian intelligence service recruited Snowden when he was still working for the NSA or even earlier? What if some other hypothetical Russian mole still inside the NSA helped him? What if he was working with the Russians unwittingly, manipulated by a handler pretending to be a “hactivist” interested in Internet privacy?

In this way, How America Lost Its Secrets plunges down rabbit holes, each leading to its own Wonderland. In building up his scenarios, Epstein deploys dozens of instances of variants of the words “presume,” “assume,” and “might have.” He describes things he believes “could have been,” things he interprets as “possible,” things he supposes were “likely,” and things he maintains were “suggested.” He piles inferences atop other inferences, as with “if so, it seems plausible to believe”; “if that is the case, then”; and “if so, it wasn’t much of a leap to assume.” He weaves cobwebs of conjecture that start with phrases like “it doesn’t take a great stretch of the imagination to conclude” and “it is not difficult to imagine.”


For Epstein’s book to have value—for it to be worth reading, not just an object intelligence hard-liners might display on their shelves as a sign of their contempt for Snowden—the facts he selects to anchor and discipline his scenario-building cannot be flimsy or cherry-picked to fit his preexisting beliefs. This is important because he clearly decided early that everything pointed in the direction of the Snowden saga being a foreign espionage plot. In June 2013, as the world was still absorbing the first revelations, Epstein published a column in The Wall Street Journal asking, “Who, if anyone, aided and abetted this well-planned theft of US secrets?”2 And in May and June of 2014, he published two more columns laying out the case that “far from being a whistleblower, Mr. Snowden was a participant in an espionage operation and most likely steered from the beginning toward his massive theft, whether he knew this at first or not.”3 Given this predisposition, it is unfortunate that Epstein builds his imagined scenarios upon allegations that may not be real facts.

For example, Epstein gives sinister significance to the “fact” that Snowden arrived in Hong Kong eleven days before he checked into the hotel where he met the journalists, leaving his activities during that period a mystery. Snowden has insisted that he was in that hotel the whole time, waiting for the journalists to arrive. In one of his columns written in 2014, Epstein first claimed that there was an eleven-day mystery gap, citing his conversation with an unnamed hotel security guard. I am aware of no independent verification of this allegation. So as things stand, this “fact” appears to be vaporous.

Epstein also makes important factual omissions, in places even overlooking crucial information that he had mentioned elsewhere. For example, laying out the case that Snowden may have decided to concoct a whistle-blower cover story at some point after he had already started copying documents for some other purpose, Epstein stresses that Snowden’s most famous leaked document—a classified intelligence court order requiring Verizon to turn over all its customers’ phone records, which “gave him credentials as a whistle-blower”—was issued in April 2013, yet Snowden had been copying files since 2012. But other documents described the program for collecting bulk domestic phone records, including a classified inspector general report Snowden also leaked; eighty-seven pages earlier, Epstein had noted that Snowden read that report in 2012.

It would be eye-glazing to compile a comprehensive list of Epstein’s doubtful “facts,” but one more is worth scrutinizing because Epstein hangs such heavy weight on it: the allegation that Snowden brought files with him to Russia, despite his denials. A Hong Kong lawyer who represented Snowden has publicly said he witnessed Snowden destroy his hard drives before leaving that city; Epstein interviewed the Hong Kong lawyer, but does not mention this corroboration. Instead, he focuses on a brief exchange during a September 2013 interview of Snowden’s Russian lawyer: the interviewer asked, “So he does have some materials that haven’t been made public yet?” and the Russian lawyer replied, “Certainly.”

For his book research, Epstein says he asked the Russian lawyer about that interview, which was conducted in Russian but translated into English before being broadcast and published, and whether the exchange was accurate. The lawyer affirmed that it was. Based on this, Epstein repeatedly states that the Russian lawyer disclosed that Snowden brought documents to Moscow; once he even embellishes it, writing that in this exchange the Moscow lawyer had disclosed that Snowden still had access in Russia to additional files that he had not given to the journalists in Hong Kong.

Yet the interview transcript shows that this exchange was ambiguous. The context, which Epstein omits, was a discussion of how the ongoing publication of new articles citing Snowden’s leaks did not mean that he was still making new leaks from Russia; rather the journalists were still just working through files he had given them in Hong Kong. So maybe this was a garbled conversational moment, and the Russian lawyer was saying that the journalists had still more unpublished materials to work with. Or maybe, in that 2013 interview, he was just playing along to gin up intrigue.

For that matter, when the lawyer later told Epstein that it was accurate, was he merely affirming the English translation of his 2013 words, or did he understand himself to be confirming the interpretive gloss Epstein placed on them? It seems to me that a journalist who wanted to know the truth, even at the risk of undermining his book project, would have followed up by asking the lawyer to clarify explicitly whether he was saying that Snowden had brought files with him to Russia and, if so, how the lawyer knew that he had done so and how he accounted for his client saying otherwise. By Epstein’s account, after obtaining this murky confirmation, he instead changed the subject. That left him free to construe this exchange as having generated a “fact” consistent with his thesis.

There is a related problem. Epstein gets many facts about surveillance issues wrong, calling into question his competence to serve as a guide to thinking seriously about the Snowden saga. He gets dates wrong, calls an important technology by the wrong name, and inaccurately describes various programs and a presidential directive Snowden leaked. His botched discussion of the Prism system, which Snowden disclosed, is a troubling example. The government uses Prism to collect from American webmail providers like Gmail, without a warrant, the e-mails of noncitizens abroad whose accounts have been targeted by intelligence officials for surveillance. When Americans communicate with those targets, the government also “incidentally” gathers those Americans’ e-mails to and from the target without a warrant. Epstein reassures his readers three times that every few months, the NSA sifts through all the e-mails it has gathered via Prism in order to filter out and purge “whatever information was accidentally picked up about Americans.” That is a fake fact.

In reality, the NSA does not filter out Americans’ messages gathered via Prism. Indeed, it shares raw messages gathered via the Prism system with the Central Intelligence Agency, the Federal Bureau of Investigation, and the National Counterterrorism Center. Once-secret rules permit officials at all those agencies to search that trove for intelligence purposes using the names of Americans and to read any private e-mails they find. FBI agents may also do so when investigating ordinary criminal suspects. When Congress in 2017 extends the law that authorizes Prism, reformers are hoping to close this so-called “backdoor search loophole” by requiring warrants to search for Americans’ e-mails within the Prism trove. Because this policy debate is attributable to Snowden’s leaks, Epstein’s misinformation about Prism is no small detail.

Edward Snowden seen live from Moscow at the Norwegian PEN event ‘Waiting for Snowden,’ where he was awarded the Ossietzky Prize for ‘outstanding contributions to freedom of expression,’ Oslo, November 2016

Berit Roald/AFP/Getty Images

Edward Snowden seen live from Moscow at the Norwegian PEN event ‘Waiting for Snowden,’ where he was awarded the Ossietzky Prize for ‘outstanding contributions to freedom of expression,’ Oslo, November 2016


Epstein argues that views differ about Snowden because the public and the media lack good information, accepting what Snowden says at face value and omitting whatever does not fit that narrative because of their “confirmation bias.” By contrast, he writes, those who hold darker views about Snowden include lawmakers and officials who “base it on classified reports” and “have been at least partially briefed” about the NSA’s investigation. Here he cites several of the latter group who said Snowden’s leaks were damaging and unjustified, including two who said in 2014 that they thought he must be a spy, although Epstein only names one of those two. But Epstein omits what Chris Inglis, who was deputy director of the NSA from 2006 to 2014 and oversaw that investigation, said last March when asked whether Snowden had acted as a spy or from his own convictions:

Here is what I surmise based upon a careful observation of the facts available to me. It does seem clear that his intention was to go to Latin or South America after he revealed all of this material in Hong Kong. He worked very hard and his lawyers worked very hard on his behalf to actually achieve that in the days and weeks afterwards…. I don’t think that he was in the employ of the Chinese or the Russians. I don’t see any evidence that would indicate that. And even if they are careful in terms of practicing denial and deception, I think there would be certain tell-tales….

Epstein also says little about Snowden’s comments criticizing Russia’s Internet policies and human rights record.4 But those comments have heightened chatter about what will happen to him under the Trump administration: Might Vladimir Putin extradite him to the United States as a gift or a bargaining chip? In a recent interview, Snowden said he found such talk perversely encouraging, since nations do not trade away their spies.

The premise of this chatter dovetails with an odd twist at the conclusion of Epstein’s book. Without much warning, he writes that he sees “no reason to doubt [Snowden’s] explanation that he stole NSA documents to expose its surveillance because he believed that it was an illicit intrusion into the privacy of individuals.” Epstein continues to criticize Snowden for taking documents that did not concern “domestic” spying, and he still maintains, vaguely, that by the end Snowden’s “mission evolved, deliberately or not, into one that led him to disclose key communications intelligence secrets to a foreign power.” But he states that he “fully” accepts that Snowden “began as a whistle-blower, not as a spy,” and was still acting as a whistle-blower when he reached out to the journalists.

By pulling back at the end of his book, Epstein tries to have it both ways: weaving conspiracy theories while maintaining plausible deniability and some veneer of evidence-based journalism. But his indulgence in speculation, his treatment of questionable claims as established facts, and his misunderstanding of surveillance combine to undermine his book’s credibility. How America Lost Its Secrets fails to live up to Epstein’s own principle, jotted down on that 3×5 card for his debate with Oliver Stone about JFK so many years ago: when a nonfiction writer reaches the limits of discoverable fact, he is supposed to stop—not fill in whatever gaps exist with his imagination, no matter how strong his intuition or suspicion.