Three days ago, the Washington Post reported the results of an investigation into a large collection of files provided by Edward Snowden. Reviewing 160,000 intercepted electronic conversations and 8,000 other documents, which Snowden apparently accessed on NSA servers after that agency collected them, the Post’s reporters found that nearly half of them contained information pertaining to U.S. citizens. Overall, the article says, the sample showed that the government scooped up information on nine bystanders (as it were) for every “targeted” individual under electronic surveillance. On that basis, the reporters speculate that the NSA may have collected information on as many as 800,000 non-target individuals in 2013.

I don’t intend to comment here on the legality, ethics, or wisdom of the NSA’s programs or the Snowden leaks. But I do think this report is fascinating and important. And I think it’s worth considering from the standpoint of digital history.

It seems to me that the Post report is a description of a cutting-edge form of historical methodology. Rightly or wrongly—and this is an idea I find pretty disquieting—the U.S. intelligence community seems to be engaged in collecting much the same information that a historian with similar tools would be.

Consider one example of the information the Post says it found in that document dump. One intelligence summary, the article says, “makes fun of a suspected kidnapper, newly arrived in Syria before the current civil war, who begs for employment as a janitor and makes wide-eyed observations about the state of undress displayed by women on local beaches.”

Now, as private and mundane as it seems, that sort of information could certainly be a legitimate part of a criminal investigation. But I’m struck by something else: how much that intelligence summary sounds like what historians would go out of their way to collect. (It sounds, for example, very much like the sort of thing historians are interested in when they write about Sayyid Qutb’s experiences in Colorado.) The information relates not to guilt or tactics or even immediate motives, but to something more basic: the subject’s worldview, his participation in some larger world of meaning.

Or consider this story of ambivalence, conflicting desires, and feelings of guilt. In the Post’s words:

She was 29 and shattered by divorce, converting to Islam in search of comfort and love. He was three years younger, rugged and restless. … One day when she was sick in bed, he brought her tea. Their faith forbade what happened next, and later she recalled it with shame. … Still, a romance grew. They fought. They spoke of marriage. They fought again. All of this was in the files because, around the same time, he went looking for the Taliban. … On May 30, 2012, without a word to her, he boarded a plane to begin a journey to Kandahar. He left word that he would not see her again.

Now, that’s fantastic history. It’s also, perhaps, an important piece of contemporary sociology. It could appear in a RAND report, a Jon Krakauer book, or some yet-unwritten academic history of our times. (The prose quality would vary, of course.) According to the Washington Post, it’s history drawn from 800 pages of conversation transcripts collected by the NSA and the Australian Signals Directorate—a detailed and intimate documentary collection of the sort most historians can only dream about.

Of course, most historians and sociologists don’t have the ability to have their research subjects assassinated.

Still, the similarity makes me worry about historical ethics in a digital age.

At first glance, there’s no problem at all. We don’t complain when historians read dead people’s mail; why should we worry about whether historians can read people’s email or chat logs? And we’ve always had a desire to peel back any layers of deceit and obfuscation that might hide forms of truth about the past. But consider.

For most of us, the vast majority of the documents and artifacts we review are thoroughly public sources, however intimate they may be in content. I don’t mean we’re using government documents. I mean we’re relying primarily on sources created, preserved, gifted, and cataloged with some kind of public performance in mind.

We’ve probably all had the experience of reading “private” letters that we knew were always meant for many eyes. We’ve studied the provenance of manuscripts and furnishings that were handed down through families, then gifted to museums with conditions attached. Even when we document (for example) illicit sexual or business relationships or other embarrassing activities, we often do it with the needs and sensitivities of living descendants and kinship groups in mind, depending on how far back our research takes us. And we’re not grave-robbers. When someone violates the unwritten rules of this kind of investigation—for instance, by prying into a historical figure’s sexual habits without any apparent motive except malice or ideological opposition—there’s sometimes at least a sense among her peers that the scholar is behaving in an unprofessional way.

What happens to history when the boundary between publicity and record-keeping disappears?

Today, private corporations as well as governments can collect vast treasuries of information about the minute-by-minute lives of millions of private individuals. They can track what we look at, what we buy (offline as well as online), where we (or at least our phones and cars and debit cards) go, what we put in our calendars, with whom we’ve been photographed, what we “like,” what makes us happy and sad, how much we exercise, and sometimes even what we write and then decide not to show anyone. These are not forms of record-keeping that can happen; these are all forms of record-keeping that, according to recent reports, do happen. Of course, these digital records may be lost long before historians get to them. But what are the responsibilities of historians if information on that scale becomes available to us, having escaped the control not only of its immediate subjects but also of their families and other relations?

Most of us agree that it can be good and proper to know something about a Founder’s sex life. But would it be good and proper for future historians to know the late-night web-browsing habits of millions of private but identifiable 21st-century citizens? We can agree that the bodily health of a dead U.S. president can be worth studying. But what about the medical files of named, traceable private citizens? We like to pore over records of criminal cases, which preserve stories about formerly private people who were thrust into the public gaze. But what if we someday can go trawling through files looking for unreported crimes and abuses that we consider historically significant?

It seems to me that information on such a scale, in which the boundary between private and public (or between private and really really private) record-keeping is entirely erased, may call for professional moral commitments that most of us don’t have to make today. Offhand, I’m thinking about some of the ethical standards attached to various forms of anthropological research methods. If so, I’m not sure the historical profession has the traditions or philosophical tools it will need to formulate those standards on its own.